SlideShare a Scribd company logo

Building Enterprise Apps for Big Data with Cascading

Paco Nathan, profile picture
Paco Nathan

The document discusses the development of enterprise applications for big data using the Cascading API, highlighting the evolution of data processing from traditional methods to modern machine learning and analytics. It emphasizes the importance of forming interdisciplinary data science teams, understanding data access patterns, and adapting to new technologies to extract ROI from big data efficiently. Case studies and use cases illustrate best practices in applying these concepts across various industries.

Technology
Related topics:
Data Science Insights
1 of 77
Downloaded 158 times
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
Building Enterprise Apps for Big Data with Cascading Paco Nathan Document Collection Scrub Tokenize token Concurrent, Inc. M HashJoin Regex Left token GroupBy R Stop Word token List RHS pnathan@concurrentinc.com Count @pacoid Word Count Copyright @2012, Concurrent, Inc.
Enterprise Apps for Big Data with Cascading 1. backstory: how we got here 2. build: Data Science teams 3. pattern: common use cases 4. intro: Cascading API 5. tutorial: for the impatient 6. code: sample apps
Intro to Cascading Document Collection Scrub Tokenize token M HashJoin Regex Left token GroupBy R Stop Word token List RHS Count Word Count 1. backstory: how we got here
inflection point huge Internet successes after 1997 holiday season… 1997 AMZN, EBAY, Inktomi (YHOO Search), then GOOG 1998 consider this metric: annual revenue per customer / amount of data stored which dropped 100x within a few years after 1997 2004 storage and processing costs plummeted, now we must work much smarter to extract ROI from Big Data… our methods must adapt “conventional wisdom” of RDBMS and BI tools became less viable; however, business cadre was still focused on pivot tables and pie charts… which tends toward inertia! MapReduce and the Hadoop open source stack grew directly out of that contention… however, that effort ＋ only solves parts of the puzzle
inflection point: consequences Geoffrey Moore (Mohr Davidow Ventures, author of Crossing The Chasm) Hadoop Summit, 2012: “All of Fortune 500 is now on notice over the next 10-year period.” Amazon and Google as exemplars of massive disruption in retail, advertising, etc. data as the major force displacing Global 1000 over the next decade, mostly through apps — verticals, leveraging domain expertise Michael Stonebraker (INGRES, PostgreSQL,Vertica,VoltDB, etc.) XLDB, 2012: “Complex analytics workloads are now displacing SQL as the basis  for Enterprise apps.”
primary sources Amazon “Early Amazon: Splitting the website” – Greg Linden glinden.blogspot.com/2006/02/early-amazon-splitting-website.html eBay “The eBay Architecture” – Randy Shoup, Dan Pritchett addsimplicity.com/adding_simplicity_an_engi/2006/11/you_scaled_your.html addsimplicity.com.nyud.net:8080/downloads/eBaySDForum2006-11-29.pdf Inktomi (YHOO Search) “Inktomi’s Wild Ride” – Erik Brewer (0:05:31 ff) youtube.com/watch?v=E91oEn1bnXM Google “The Birth of Google” – John Battelle wired.com/wired/archive/13.08/battelle.html “Underneath the Covers at Google” – Jeff Dean (0:06:54 ff) youtube.com/watch?v=qsan-GQaeyk perspectives.mvdirona.com/2008/06/11/JeffDeanOnGoogleInfrastructure.aspx
the world before… BI, SQL, and highly optimized code
data innovation: circa 1996 Stakeholder Customers Excel pivot tables PowerPoint slide decks strategy BI Product Analysts requirements SQL Query optimized Engineering code Web App result sets transactions RDBMS
the world after… machine learning, leveraging log files
data innovation: circa 2001 Stakeholder Product Customers dashboards UX Engineering models servlets recommenders Algorithmic + Web Apps Modeling classifiers Middleware aggregation event SQL Query history result sets customer transactions Logs DW ETL RDBMS
the world ahead… what our customers are doing now
data innovation: circa 2013 Customers Data Apps business Domain process Workflow Prod Expert dashboard Web Apps, metrics History services Mobile, data etc. s/w science dev Data Planner Scientist social discovery optimized interactions + capacity transactions, Eng endpoints modeling content App Dev Data Access Patterns Hadoop, Log In-Memory etc. Events Data Grid Ops DW Ops batch "real time" Cluster Scheduler introduced existing capability SDLC RDBMS RDBMS
a key difference…
statistical thinking Process Variation Data Tools employing a mode of thought which includes both logical and analytical reasoning: evaluating the whole of a problem, as well as its component parts; attempting to assess the effects of changing one or more variables this approach attempts to understand not just problems and solutions, but also the processes involved and their variances particularly valuable in Big Data work when combined with hands-on experience in physics – roughly 50% of my peers come from physics or physical engineering… programmers typically don’t think this way… however, both systems engineers and data scientists must!
reference by Leo Breiman Statistical Modeling: The Two Cultures Statistical Science, 2001 bit.ly/eUTh9L
Intro to Cascading Document Collection Scrub Tokenize token M HashJoin Regex Left token GroupBy R Stop Word token List RHS Count Word Count 2. build: Data Science teams
core values Data Science teams develop actionable insights, building confidence for decisions that work may influence a few decisions worth billions (e.g., M&A) or billions of small decisions (e.g., AdWords) probably somewhere in-between… Wikipedia solving for pattern, at scale. an interdisciplinary pursuit which requires teams, not sole players
most valuable skills approximately 80% of the costs for data-related projects get spent on data preparation – mostly on cleaning up data quality issues: ETL, log file analysis, etc. unfortunately, data-related budgets for many companies tend to go into frameworks which can only be used after clean up most valuable skills: ‣ learn to use programmable tools that prepare data ‣ learn to generate compelling data visualizations ‣ learn to estimate the confidence for reported results ‣ learn to automate work, making analysis repeatable D3 the rest of the skills – modeling, algorithms, etc. – those are secondary
social caveats “This data cannot be correct!” may be an early warning about an organization itself much depends on how the people whom you work alongside tend to arrive at decisions: ‣ probably good: Induction, Abduction, Circumscription ‣ probably poor: Deduction, Speculation, Justification in general, one good data visualization puts many ongoing verbal arguments to rest however, let domain experts handle “data storytelling”, not data scientists xkcd
the science in data science? edoMpUsserD:IUN tcudorP ylppA lenaP yrotnevnI tneilC tcudorP evomeR lenaP yrotnevnI tneilC in a nutshell, what we do… edoMmooRyM:IUN edoMmooRcilbuP:IUN ydduB ddA nigoL etisbeW vd edoMsdneirF:IUN edoMtahC:IUN egasseM a evaeL G1 :gniniamer ecaps sserddA dekcilCeliforPyM:IUN edoMstiderCyuB:IUN tohspanS a ekaT egapemoH nwO tisiV elbbuB a epyT taeS egnahC wodniW D3 nepO dneirF ddA revO tcudorP pilF lenaP yrotnevnI tneilC lenaP tidE ‣ estimate probability woN tahC teP yalP teP deeF 2 petS egaP traC esahcruP edaM remotsuC M215 :gniniamer ecaps sserddA gnihtolC no tuP bew :metI na yuB edoMeivoM:IUN ytinummoc ,tneilc :detratS weiV eivoM teP weN etaerC detrats etius tset :tseTytivitcennoC emag pazyeh dehcnuaL eciov mooRcilbuP tahC egasseM yadhtriB edoMlairotuT:IUN ybbol semag dehcnuaL noitartsigeR euqinU ‣ calculate analytic variance edoMpUsserD:IUN tcudorP ylppA lenaP yrotnevnI tneilC tcudorP evomeR lenaP yrotnevnI tneilC edoMmooRyM:IUN edoMmooRcilbuP:IUN ydduB ddA nigoL etisbeW vd edoMsdneirF:IUN edoMtahC:IUN egasseM a evaeL G1 :gniniamer ecaps sserddA dekcilCeliforPyM:IUN edoMstiderCyuB:IUN tohspanS a ekaT egapemoH nwO tisiV elbbuB a epyT taeS egnahC dneirF ddA revO tcudorP pilF lenaP yrotnevnI tneilC lenaP tidE woN tahC teP yalP teP deeF 2 petS egaP traC esahcruP edaM remotsuC M215 :gniniamer ecaps sserddA gnihtolC no tuP bew :metI na yuB edoMeivoM:IUN ytinummoc ,tneilc :detratS weiV eivoM teP weN etaerC detrats etius tset :tseTytivitcennoC emag pazyeh dehcnuaL eciov mooRcilbuP tahC egasseM yadhtriB edoMlairotuT:IUN ybbol semag dehcnuaL noitartsigeR euqinU wodniW D3 nepO ‣ manipulate order complexity ‣ make use of learning theory + collab with DevOps, Stakeholders + reduce our work to cron entries
synthesis of the above MapReduce is Good Enough? Jimmy Lin, U Maryland + Twitter arxiv.org/pdf/1209.2191v1.pdf A Few Useful Things to Know about Machine Learning Pedro Domingos, U Washington homes.cs.washington.edu/~pedrod/papers/cacm12.pdf
team process = needs help people ask the discovery right questions allow automation to place modeling informed bets deliver products at integration scale to customers build smarts into apps product features Gephi keep infrastructure systems running, cost-effective
team composition = roles Domain Expert business process, stakeholder data science Data data prep, discovery, Scientist modeling, etc. Document Collection Scrub Tokenize token M HashJoin Regex Left token GroupBy R Stop Word token List RHS App Dev software engineering, Count automation Word Count Ops systems engineering, access introduced capability
matrix = needs × roles nn o overy very elliing e ng ratiio rat o apps apps tem tem ss diisc d sc mod mod nteg ii nteg sys sys stakeholder scientist developer ops
matrix: example team nn o overy very elliing e ng ratiio rat o apps apps tem tem ss diisc d sc mod mod nteg ii nteg sys sys stakeholder scientist developer ops summary: this team seems heavy on systems, may need more overlap between modeling and integration, particularly among team leads
Q: Can I simply hire one rockstar data scientist to cover all this work?
A: No, interdisciplinary work requires teams. A: Hire leads who speak the lingo of each domain. A: Hire people who cover 2+ roles, when possible.
reference by DJ Patil Data Jujitsu O’Reilly, 2012 amazon.com/dp/B008HMN5BE Building Data Science Teams O’Reilly, 2011 amazon.com/dp/B005O4U3ZE
Intro to Cascading Document Collection Scrub Tokenize token M HashJoin Regex Left token GroupBy R Stop Word token List RHS Count Word Count 3. pattern: common use cases
CAP theorem purpose: theoretical limits for data access patterns essence: ‣ consistency ‣ availability ‣ partition tolerance best case scenario: you may pick two … or spend billions struggling to obtain all three at scale (GOOG) translated: cost of doing business www.cs.berkeley.edu/~brewer/cs262b-2004/PODC-keynote.pdf julianbrowne.com/article/viewer/brewers-cap-theorem
data access patterns design patterns: originated in consensus negotiation for architecture, later used in software engineering consider the corollaries in large-scale data work… essence: select data frameworks based on your data access patterns in other words, decouple use cases based on needs – avoid the “one size fits all” (OSFA) anti-pattern let’s review some examples…
access → frameworks → forfeits financial transactions general ledger in RDBMS CAx ad-hoc queries RDS (hosted MySQL) CAx reporting, dashboards like Pentaho CAx log rotation/persistence like Riak xxP search indexes like Lucene/Solr xAP static content, archives S3 (durable storage) xAP customer facts like Redis, Membase xAP distributed counters, locks, sets like Redis x A P* data objects CRUD key/value – like, NoSQL on MySQL CxP authoritative metadata like Zookeeper CxP data prep, modeling at scale like Hadoop/Cascading + R CxP graph analysis like Hadoop + Redis + Gephi CxP data marts like Hadoop/HBase CxP
access → frameworks → forfeits financial transactions general ledger in RDBMS CAx ad-hoc queries RDS (hosted MySQL) CAx reporting, dashboards like Pentaho CAx log rotation/persistence like Riak xxP search indexes like Lucene/Solr xAP static content, archives S3 (durable storage) xAP customer facts like Redis, Membase xAP distributed counters, locks, sets like Redis x A P* data objects CRUD key/value – like, NoSQL on MySQL CxP authoritative metadata like Zookeeper CxP data prep, modeling at scale like Hadoop/Cascading + R CxP graph analysis like Hadoop + Redis + Gephi CxP data marts like Hadoop/HBase CxP
and, since “One Size Fits All”… doesn’t
a selection of great tools… reporting: visualization: Graphite, PowerPivot, ggplot2, D3, Gephi analytics/modeling: Pentaho, Jaspersoft, SAS R, Weka, Matlab, PMML, GLPK text: LDA, WordNet, OpenNLP, Mallet, Bixo, NLTK apps: Cascading, Scalding, Cascalog, R markdown, SWF scale-out: Scalr, RightScale, CycleComputing, vFabric, Beanstalk graph: column: Gremlin, Vertica, GraphLab, HBase, key/val: index: relational: Neo4J Drill, Redis, Lucene/Solr, usual suspects Dynamo Membase, ElasticSearch MySQL imdg: Spark, Storm, hadoop: EMR, HW, MapR, machine data: Gigaspaces EMC, Azure, Compute Splunk, collectd, durable storage: Nagios S3, ASV, GCS, Riak, Couch
common use cases app patterns
use case: marketing funnel • must optimize a very large ad spend • different vendors report different metrics Wikipedia • seasonal variation distorts performance • some campaigns are much smaller than others • hard to predict ROI for incremental spend approach: • log aggregation, followed with cohort analysis • bayesian point estimates compare different-sized ad tests • customer lifetime value quantifies ROI of new leads • time series analysis normalizes for seasonal variation • geolocation adjusts for regional cost/benefit • linear programming models estimate elasticity of demand
use case: ecommerce fraud • sparse data means lots of missing values stat.berkeley.edu • “needle in a haystack” lack of training cases • answers are available in large-scale batch, results are needed in real-time event processing • not just one pattern to detect – many, ever-changing approach: • random forest (RF) classifiers predict likely fraud • subsampled data to re-balance training sets • impute missing values based on density functions • train on massive log files, run on in-memory grid • adjust metrics to minimize customer support costs • detect novelty – report anomalies via notifications
use case: customer segmentation • many millions of customers, hard to determine which features resonate Mathworks • multi-modal distributions get obscured by the practice of calculating an “average” • not much is known about individual customers approach: • connected components for sessionization, determining uniques from logs • estimates for age, gender, income, geo, etc. • clustering algorithms to group into market segments • social graph infers “unknown” relationships • covariance/heat maps visualizes segments vs. feature sets
use case: monetizing content • need to suggest relevant content which would Digital Humanities otherwise get buried in the back catalog • big disconnect between inventory and limited performance ad market • enormous amounts of text, hard to categorize approach: • text analytics glean key phrases from documents • hierarchical clustering of char frequencies detects lang • latent dirichlet allocation (LDA) reduces dimension to topic models • recommenders suggest similar topics to customers • collaborative filters connect known users with less known
Intro to Cascading Document Collection Scrub Tokenize token M HashJoin Regex Left token GroupBy R Stop Word token List RHS Count Word Count 4. intro: Cascading API
Cascading API: purpose ‣ simplify data processing development and deployment ‣ improve application developer productivity ‣ enable data processing application manageability
Cascading API: a few facts Java open source project (ASL 2) using Git, Gradle, Maven, JUnit, etc. in production (~5 yrs) at hundreds of enterprise Hadoop deployments: Finance, Health Care, Transportation, other verticals studies published about large use cases: Twitter, Etsy, Airbnb, Square, Climate Corporation, FlightCaster, Williams-Sonoma partnerships and distribution with SpringSource, Amazon AWS, Microsoft Azure, Hortonworks, MapR, EMC several open source projects built atop, managed by Twitter, Etsy, etc., which provide substantial Machine Learning libraries DSLs available in Scala, Clojure, Python (Jython), Ruby (JRuby), Groovy data “taps” integrate popular data frameworks via JDBC, Memcached, HBase, plus serialization in Apache Thrift, Avro, Kyro, etc. entire app compiles into a single JAR: fully connected for compiler optimization, exception handling, debugging, config, scheduling, etc.
Cascading API: a few quotes “Cascading gives Java developers the ability to build Big Data applications on Hadoop using their existing skillset … Management can really go out and build a team around folks that are already very experienced with Java. Switching over to this is really a very short exercise.” CIO, Thor Olavsrud, 2012-06-06 cio.com/article/707782/Ease_Big_Data_Hiring_Pain_With_Cascading “Masks the complexity of MapReduce, simplifies the programming, and speeds you on your journey toward actionable analytics … A vast improvement over native MapReduce functions or Pig UDFs.” 2012 BOSSIE Awards, James Borck, 2012-09-18 infoworld.com/slideshow/65089 “Company’s promise to application developers is an opportunity to build and test applications on their desktops in the language of choice with familiar constructs and reusable components” Dr. Dobb’s, Adrian Bridgwater, 2012-06-08 drdobbs.com/jvm/where-does-big-data-go-to-get-data-inten/240001759
data+code “political spectrum” “Notes from the Mystery Machine Bus” by Steve Yegge, Google goo.gl/SeRZa “conservative” “liberal” (mostly) Enterprise (mostly) Start-Up risk management customer experiments assurance flexibility well-defined schema schema follows code explicit configuration convention type-checking compiler interpreted scripts wants no surprises wants no impediments Java, Scala, Clojure, etc. PHP, Ruby, Python, etc. Cascading, Scalding, Cascalog, etc. Hive, Pig, Hadoop Streaming, etc.
Cascading API: adoption As Enterprise apps move into Hadoop and related BigData frameworks, risk profiles shift toward more conservative programming practices Cascading provides a popular API for defining and managing Enterprise data workflows
enterprise data workflows Tuples, Pipelines, Endpoints, Operations, Joins, Assertions, Traps, etc. …in other words, “plumbing” Document Collection Scrub Tokenize token M HashJoin Regex Left token GroupBy R Stop Word token List RHS Count Word Count
data workflows: team ‣ Business Stakeholder POV: business process management for workflow orchestration (think BPM/BPEL) ‣ Systems Integrator POV: system integration of heterogenous data sources and compute platforms ‣ Data Scientist POV: a directed, acyclic graph (DAG) on which we can apply Amdahl's Law, etc. ‣ Data Architect POV: a physical plan for large-scale data flow management ‣ Software Architect POV: a pattern language, similar to plumbing or circuit design Document Collection ‣ App Developer POV: M Tokenize Scrub token API bindings for Java, Scala, Clojure, Jython, JRuby, etc. Stop Word List HashJoin Left RHS Regex token GroupBy token R Count ‣ Systems Engineer POV: Word Count a JAR file, has passed CI, available in a Maven repo
data workflows: layers business domain expertise, business trade-offs, process operating parameters, market position, etc. API Java, Scala, Clojure, Jython, JRuby, Groovy, etc. language …envision whatever runs in a JVM optimize / schedule major changes in technology now Document Collection Scrub Tokenize token physical M HashJoin Regex Left token GroupBy R plan Stop Word token List RHS Count Word Count compute Apache Hadoop, in-memory local mode “assembler” code substrate …envision GPUs, streaming, etc. machine data Splunk, Nagios, Collectd, New Relic, etc.
data workflows: SQL Relational SQL parser logical plan, optimized based on stats physical plan query history, table stats b-trees, etc. ERD table schema catalog
data workflows: SQL vs. JVM Relational Cascading + Driven SQL parser SQL-92 compliant parser (in progress) logical plan, TODO: logical plan, optimized based on stats optimized based on stats physical plan API “plumbing” query history, app history, table stats tuple stats b-trees, etc. distributed compute substrate: Hadoop, in-memory, etc. ERD flow diagram table schema tuple schema catalog endpoint usage DB
Intro to Cascading Document Collection Scrub Tokenize token M HashJoin Regex Left token GroupBy R Stop Word token List RHS Count Word Count 5. tutorial: for the impatient
“Cascading for the Impatient” cascading.org/category/impatient/ ‣ a series of introductory tutorials and code samples ‣ 1:1 code comparisons in Scalding, Cascalog, Pig, Hive Document Collection Scrub Tokenize token M HashJoin Regex Left token GroupBy R Stop Word token List RHS Count Word Count
1: copy public class   Main   {   public static void   main( String[] args )     {     String inPath = args[ 0 ];     String outPath = args[ 1 ]; Source     Properties props = new Properties();     AppProps.setApplicationJarClass( props, Main.class );     HadoopFlowConnector flowConnector = new HadoopFlowConnector( props );     // create the source tap     Tap inTap = new Hfs( new TextDelimited( true, "t" ), inPath ); M     // create the sink tap     Tap outTap = new Hfs( new TextDelimited( true, "t" ), outPath ); Sink     // specify a pipe to connect the taps     Pipe copyPipe = new Pipe( "copy" );     // connect the taps, pipes, etc., into a flow     FlowDef flowDef = FlowDef.flowDef().setName( "copy" )      .addSource( copyPipe, inTap )      .addTailSink( copyPipe, outTap );     // run the flow     flowConnector.connect( flowDef ).complete(); 1 mapper     }   } 0 reducers 10 lines code
wait! ten lines of code for a file copy… seems like a lot.
same JAR, any scale… MegaCorp Enterprise IT: Pb’s data 1000+ node private cluster EVP calls you when app fails runtime: days+ Production Cluster: Tb’s data EMR w/ 50 HPC Instances Ops monitors results runtime: hours – days Staging Cluster: Gb’s data EMR + 4 Spot Instances CI shows red or green lights runtime: minutes – hours Your Laptop: Mb’s data Hadoop standalone mode passes unit tests, or not runtime: seconds – minutes
2: word count Document Collection Tokenize GroupBy M token Count R Word Count 1 mapper 1 reducer 18 lines code gist.github.com/3900702
Cascading / Java Document Collection M Tokenize GroupBy token Count String docPath = args[ 0 ]; R Word String wcPath = args[ 1 ]; Count Properties properties = new Properties(); AppProps.setApplicationJarClass( properties, Main.class ); HadoopFlowConnector flowConnector = new HadoopFlowConnector( properties ); // create source and sink taps Tap docTap = new Hfs( new TextDelimited( true, "t" ), docPath ); Tap wcTap = new Hfs( new TextDelimited( true, "t" ), wcPath ); // specify a regex to split "document" text lines into token stream Fields token = new Fields( "token" ); Fields text = new Fields( "text" ); RegexSplitGenerator splitter = new RegexSplitGenerator( token, "[ [](),.]" ); // only returns "token" Pipe docPipe = new Each( "token", text, splitter, Fields.RESULTS ); // determine the word counts Pipe wcPipe = new Pipe( "wc", docPipe ); wcPipe = new GroupBy( wcPipe, token ); wcPipe = new Every( wcPipe, Fields.ALL, new Count(), Fields.ALL ); // connect the taps, pipes, etc., into a flow FlowDef flowDef = FlowDef.flowDef().setName( "wc" ) .addSource( docPipe, docTap )  .addTailSink( wcPipe, wcTap ); // write a DOT file and run the flow Flow wcFlow = flowConnector.connect( flowDef ); wcFlow.writeDOT( "dot/wc.dot" ); wcFlow.complete();
Scalding / Scala Document Collection M Tokenize GroupBy token Count R Word Count // Sujit Pal // sujitpal.blogspot.com/2012/08/scalding-for-impatient.html package com.mycompany.impatient import com.twitter.scalding._ class Part2(args : Args) extends Job(args) {   val input = Tsv(args("input"), ('docId, 'text))   val output = Tsv(args("output"))   input.read.     flatMap('text -> 'word) { text : String => text.split("""s+""") }.     groupBy('word) { group => group.size }.     write(output) }
Cascalog / Clojure Document Collection M Tokenize GroupBy token Count R Word Count ; Paul Lam ; github.com/Quantisan/Impatient (ns impatient.core   (:use [cascalog.api]         [cascalog.more-taps :only (hfs-delimited)])   (:require [clojure.string :as s]             [cascalog.ops :as c])   (:gen-class)) (defmapcatop split [line]   "reads in a line of string and splits it by regex"   (s/split line #"[[](),.)s]+")) (defn -main [in out & args]   (?<- (hfs-delimited out)        [?word ?count]        ((hfs-delimited in :skip-header? true) _ ?line)        (split ?line :> ?word)        (c/count ?count)))
Hive Document Collection M Tokenize GroupBy token Count R Word Count -- Steve Severance -- stackoverflow.com/questions/10039949/word-count-program-in-hive CREATE TABLE input (line STRING); LOAD DATA LOCAL INPATH 'input.tsv' OVERWRITE INTO TABLE input; SELECT  word, COUNT(*) FROM input  LATERAL VIEW explode(split(text, ' ')) lTable AS word GROUP BY word ;
Pig Document Collection M Tokenize GroupBy token Count R Word Count -- kudos to Dmitriy Ryaboy docPipe = LOAD '$docPath' USING PigStorage('t', 'tagsource') AS (doc_id, text); docPipe = FILTER docPipe BY doc_id != 'doc_id'; -- specify regex to split "document" text lines into token stream tokenPipe = FOREACH docPipe GENERATE doc_id, FLATTEN(TOKENIZE(text, ' [](),.')) AS token; tokenPipe = FILTER tokenPipe BY token MATCHES 'w.*'; -- determine the word counts tokenGroups = GROUP tokenPipe BY token; wcPipe = FOREACH tokenGroups GENERATE group AS token, COUNT(tokenPipe) AS count; -- output STORE wcPipe INTO '$wcPath' USING PigStorage('t', 'tagsource'); EXPLAIN -out dot/wc_pig.dot -dot wcPipe;
3: wc + scrub Document Collection Scrub GroupBy Tokenize token token Count M R Word Count 1 mapper 1 reducer 22+10 lines code
4: wc + scrub + stop words Document Collection Scrub Tokenize token M HashJoin Regex Left token GroupBy R Stop Word token List RHS Count 1 mapper Word 1 reducer Count 28+10 lines code
5: tf-idf Unique Insert SumBy D doc_id 1 doc_id Document Collection M R M R M RHS Scrub Tokenize token HashJoin M RHS token HashJoin Regex Unique GroupBy DF Left token token token ExprFunc Count CoGroup Stop Word tf-idf List RHS M R M R M R TF-IDF M GroupBy TF doc_id, token Count GroupBy Count token M R M R Word R M R Count 11 mappers 9 reducers 65+10 lines code
6: tf-idf + tdd Unique Insert SumBy D doc_id 1 doc_id Document Collection RHS M R M R M Assert Scrub Tokenize token HashJoin Checkpoint M M RHS token HashJoin Regex Unique GroupBy DF Left token token token Count ExprFunc CoGroup tf-idf Stop Word List RHS M R M R M R TF-IDF M GroupBy TF doc_id, Failure token Count Traps GroupBy Count token M R M R Word Count R M R 12 mappers 9 reducers 76+14 lines code
deployed on AWS… elastic-mapreduce --create --name "TF-IDF" --jar s3n://temp.cascading.org/impatient/part6.jar --arg s3n://temp.cascading.org/impatient/rain.txt --arg s3n://temp.cascading.org/impatient/out/wc --arg s3n://temp.cascading.org/impatient/en.stop --arg s3n://temp.cascading.org/impatient/out/tfidf --arg s3n://temp.cascading.org/impatient/out/trap --arg s3n://temp.cascading.org/impatient/out/check aws.amazon.com/elasticmapreduce/
results? doc_id tf-idf doc02 0.9163 token air doc05 0.9163 australia doc05 0.9163 broken doc04 0.9163 california's doc04 0.9163 cause doc02 0.9163 cloudcover doc04 0.9163 death doc04 0.9163 deserts doc03 0.9163 downwind doc_id text … doc01 A rain shadow is a dry area on the lee back side of a mountainous area. doc02 0.9163 sinking doc02 This sinking, dry air produces a rain shadow, or area in the lee of a mountain doc04 0.9163 such with less rain and cloudcover. doc04 0.9163 valley doc03 A rain shadow is an area of dry land that lies on the leeward (or downwind) doc05 0.9163 women side of a mountain. doc03 0.5108 land doc04 This is known as the rain shadow effect and is the primary cause of leeward doc05 0.5108 land deserts of mountain ranges, such as California's Death Valley. doc01 0.5108 lee doc05 Two Women. Secrets. A Broken Land. [DVD Australia] doc02 0.5108 lee zoink null doc03 0.5108 leeward doc04 0.5108 leeward doc01 0.4463 area doc02 0.2231 area doc03 0.2231 area doc01 0.2231 dry doc02 0.2231 dry doc03 0.2231 dry doc02 0.2231 mountain Unique Insert SumBy D doc_id 1 doc_id Document Collection RHS M R M R M doc03 0.2231 mountain Assert Scrub Tokenize token HashJoin Checkpoint M M RHS token HashJoin Regex Unique GroupBy DF Left token doc04 0.2231 mountain token token Count ExprFunc CoGroup tf-idf Stop Word List RHS M R M R M R TF-IDF GroupBy M doc01 0.0000 rain TF doc_id, Failure token Count Traps GroupBy Count token doc02 0.0000 rain M R M R Word Count R M R doc03 0.0000 rain doc04 0.0000 rain doc01 0.0000 shadow doc02 0.0000 shadow doc03 0.0000 shadow doc04 0.0000 shadow
comparisons? compare similar code in Scalding (Scala) and Cascalog (Clojure): sujitpal.blogspot.com/2012/08/scalding-for-impatient.html based on: github.com/twitter/scalding/wiki github.com/Quantisan/Impatient based on: github.com/nathanmarz/cascalog/wiki
Intro to Cascading Document Collection Scrub Tokenize token M HashJoin Regex Left token GroupBy R Stop Word token List RHS Count Word Count 6. code: sample apps
Social Recommender filter Twitter stop words tweets calculate QA similiarity threshold min, max Neo4j LDA Redis github.com/Cascading/SampleRecommender ‣ social recommender based on Twitter: suggest users who tweet about similar stocks ‣ instead of a cross-product (potential bottleneck) this runs in parallel on Hadoop ‣ uses a stop word list to remove common words, offensive phrases, etc. ‣ one tap measures token frequency: for QA, adjust stop words, improve filter, etc. ‣ adapted in Spring by Costin Leau
SocRec: architecture Twitter filter low-freq firehose source stop words tweets batch updates ( uid, tweet, t ) checkpoint: tokenized tweets calculate checkpoint: analysis + QA similiarity token frequency curation checkpoint: similarity similar users thresholds threshold min, max sink sink sink Neo4j: social Redis graph LDA: topic results (uid: uidx, rank) trending
SocRec: results uid recommend weight carbonfiberxrm ClosingBellNews 0.1459 carbonfiberxrm DJFunkyGrrL 0.0870 ClosingBellNews DJFunkyGrrL 0.1491 CloudStocks DJFunkyGrrL 0.1206 ElmoreNicole DJFunkyGrrL 0.1798 EsNeey alexiolo_ 0.8603 ...
City of Palo Alto open data Regex Regex tree Scrub filter parser species M HashJoin Left Geohash CoPA GIS exprot Tree Metadata M RHS RHS tree Regex Checkpoint road Regex Regex tsv parser tsv filter Tree Filter GroupBy Checkpoint parser CoGroup Distance tree_dist tree_name shade M R M R M RHS M HashJoin Estimate Road Left Albedo Segments Geohash CoGroup Road Metadata GPS Failure RHS M logs Traps R road Geohash M Regex park filter reco M park github.com/Cascading/CoPA/wiki ‣ GIS export for parks, roads, trees (unstructured / open data) ‣ log files of personalized/frequented locations in Palo Alto via iPhone GPS tracks ‣ curated metadata, used to enrich the dataset ‣ could extend via mash-up with many available public data APIs Enterprise-scale app: road albedo + tree species metadata + geospatial indexing “Find a shady spot on a summer day to walk near downtown and take a call…”
CoPA: log events
CoPA: results 0.12 Estimated Tree Height (meters) 0.10 0.08 count 0 density 100 0.06 200 300 0.04 0.02 0.00 0 10 20 30 40 50 avg_height ‣ addr: 115 HAWTHORNE AVE ‣ lat/lng: 37.446, -122.168 ‣ geohash: 9q9jh0 ‣ tree: 413 site 2 ‣ species: Liquidambar styraciflua ‣ avg height 23 m ‣ road albedo: 0.12 ‣ distance: 10 m ‣ a short walk from my train stop ✔
drill-down blog, code/wiki/gists, jars, list, DevOps products: cascading.org/ github.org/Cascading/ conjars.org/ goo.gl/KQtUL concurrentinc.com/ pnathan@concurrentinc.com @pacoid

More Related Content

PDF
Enterprise Data Workflows with Cascading and Windows Azure HDInsight
Paco Nathan
 
KEY
Intro to Cascading (SpringOne2GX)
Paco Nathan
 
PDF
Hadoop and Beyond
Paco Nathan
 
DOCX
Sunshine consulting mopuru babu cv_java_j2ee_spring_bigdata_scala
Mopuru Babu
 
PDF
Platforms for data science
Deepak Singh
 
PDF
How Salesforce.com uses Hadoop
Narayan Bharadwaj
 
PDF
ScalaTo July 2019 - No more struggles with Apache Spark workloads in production
Chetan Khatri
 
PDF
Distributed Data Analysis with Hadoop and R - Strangeloop 2011
Jonathan Seidman
 
Enterprise Data Workflows with Cascading and Windows Azure HDInsight
Paco Nathan
 
Intro to Cascading (SpringOne2GX)
Paco Nathan
 
Hadoop and Beyond
Paco Nathan
 
Sunshine consulting mopuru babu cv_java_j2ee_spring_bigdata_scala
Mopuru Babu
 
Platforms for data science
Deepak Singh
 
How Salesforce.com uses Hadoop
Narayan Bharadwaj
 
ScalaTo July 2019 - No more struggles with Apache Spark workloads in production
Chetan Khatri
 
Distributed Data Analysis with Hadoop and R - Strangeloop 2011
Jonathan Seidman
 

What's hot (19)

DOCX
Anil_BigData Resume
Anil Sokhal
 
PDF
Treasure Data: Big Data Analytics on Heroku
Salesforce Developers Japan
 
DOC
Resume - Narasimha Rao B V (TCS)
Venkata Narasimha Rao B
 
PPTX
Galaxy of bits
Michal Zylinski
 
DOCX
a9TD6cbzTZotpJihekdc+w==.docx
VasimMemon4
 
PDF
Distributed Data Analysis with Hadoop and R - OSCON 2011
Jonathan Seidman
 
PDF
Advanced analytics with sap hana and r
SAP Technology
 
PDF
Data Analysis with Hadoop and Hive, ChicagoDB 2/21/2011
Jonathan Seidman
 
KEY
The Other Way of Doing Big Data
Infochimps, a CSC Big Data Business
 
DOCX
Resume
kuppani kishor
 
DOCX
PRAFUL_HADOOP
PRAFUL DASH
 
DOCX
PRAFUL_HADOOP
PRAFUL DASH
 
PDF
Treasure Data and Heroku
Treasure Data, Inc.
 
PPT
Os Lonergan
oscon2007
 
PDF
Search Engine - How to Make it
Andreas Yunanto
 
PDF
Using hadoop to expand data warehousing
DataWorks Summit
 
PPT
Hw09 Production Deep Dive With High Availability
Cloudera, Inc.
 
PDF
[Hadoop] NexR Terapot: Massive Email Archiving
Jinho Jung
 
PPTX
Steve Watt Presentation
Big Data Houston
 
Anil_BigData Resume
Anil Sokhal
 
Treasure Data: Big Data Analytics on Heroku
Salesforce Developers Japan
 
Resume - Narasimha Rao B V (TCS)
Venkata Narasimha Rao B
 
Galaxy of bits
Michal Zylinski
 
a9TD6cbzTZotpJihekdc+w==.docx
VasimMemon4
 
Distributed Data Analysis with Hadoop and R - OSCON 2011
Jonathan Seidman
 
Advanced analytics with sap hana and r
SAP Technology
 
Data Analysis with Hadoop and Hive, ChicagoDB 2/21/2011
Jonathan Seidman
 
The Other Way of Doing Big Data
Infochimps, a CSC Big Data Business
 
Resume
kuppani kishor
 
PRAFUL_HADOOP
PRAFUL DASH
 
PRAFUL_HADOOP
PRAFUL DASH
 
Treasure Data and Heroku
Treasure Data, Inc.
 
Os Lonergan
oscon2007
 
Search Engine - How to Make it
Andreas Yunanto
 
Using hadoop to expand data warehousing
DataWorks Summit
 
Hw09 Production Deep Dive With High Availability
Cloudera, Inc.
 
[Hadoop] NexR Terapot: Massive Email Archiving
Jinho Jung
 
Steve Watt Presentation
Big Data Houston
 
Ad

Viewers also liked (20)

PDF
sizeof(Object): how much memory objects take on JVMs and when this may matter
Dawid Weiss
 
PDF
Conception avec pic
nawzat
 
PPT
Introduction to programming with c,
Hossain Md Shakhawat
 
PPT
Introduction to C Programming
MOHAMAD NOH AHMAD
 
PPTX
Pic 16f877 ..
sunprass
 
PDF
Programmation des pic_en_c_part1
oussamada
 
PPT
Microcontroleur Pic16 F84
guest1e7b02
 
PDF
Cours pics16 f877
Abdellatif Mbk
 
PPT
AP Physics - Chapter 6 Powerpoint
Mrreynon
 
PPT
Ge matrix
Sunil Chichra
 
PPTX
Ge9 final ppt
Ravin Gandhi
 
PPTX
Embedded system (Chapter )
Ikhwan_Fakrudin
 
PPTX
Glasgow Coma Scale Presentation
Hayden G
 
PPT
AP Physics - Chapter 4 Powerpoint
Mrreynon
 
PPT
Basics of C programming
avikdhupar
 
ODP
Music Recommendation Tutorial
Oscar Celma
 
PDF
Test driven development in C
Amritayan Nayak
 
PDF
2015 Upload Campaigns Calendar - SlideShare
SlideShare
 
PPTX
What to Upload to SlideShare
SlideShare
 
PDF
Getting Started With SlideShare
SlideShare
 
sizeof(Object): how much memory objects take on JVMs and when this may matter
Dawid Weiss
 
Conception avec pic
nawzat
 
Introduction to programming with c,
Hossain Md Shakhawat
 
Introduction to C Programming
MOHAMAD NOH AHMAD
 
Pic 16f877 ..
sunprass
 
Programmation des pic_en_c_part1
oussamada
 
Microcontroleur Pic16 F84
guest1e7b02
 
Cours pics16 f877
Abdellatif Mbk
 
AP Physics - Chapter 6 Powerpoint
Mrreynon
 
Ge matrix
Sunil Chichra
 
Ge9 final ppt
Ravin Gandhi
 
Embedded system (Chapter )
Ikhwan_Fakrudin
 
Glasgow Coma Scale Presentation
Hayden G
 
AP Physics - Chapter 4 Powerpoint
Mrreynon
 
Basics of C programming
avikdhupar
 
Music Recommendation Tutorial
Oscar Celma
 
Test driven development in C
Amritayan Nayak
 
2015 Upload Campaigns Calendar - SlideShare
SlideShare
 
What to Upload to SlideShare
SlideShare
 
Getting Started With SlideShare
SlideShare
 
Ad

Similar to Building Enterprise Apps for Big Data with Cascading (20)

KEY
A Data Scientist And A Log File Walk Into A Bar...
Paco Nathan
 
PDF
Cascading: Enterprise Data Workflows based on Functional Programming
Paco Nathan
 
PDF
Cascading meetup #4 @ BlueKai
Paco Nathan
 
PDF
Chicago Hadoop Users Group: Enterprise Data Workflows
Paco Nathan
 
PDF
NOSQLEU - Graph Databases and Neo4j
Tobias Lindaaker
 
PDF
Pattern: an open source project for migrating predictive models onto Apache H...
Paco Nathan
 
KEY
Intro to Data Science for Enterprise Big Data
Paco Nathan
 
PPT
Search, APIs, capability management and Sensis's journey
ablebagel
 
PDF
sones company presentation
sones GmbH
 
PPTX
Anti-social Databases
William LaForest
 
PDF
Good Data: Collaborative Analytics On Demand
zsvoboda
 
PDF
Large-Scale Search Discovery Analytics with Hadoop, Mahout, Solr
DataWorks Summit
 
PPTX
Tech4Africa - Opportunities around Big Data
Steve Watt
 
PDF
hadoop @ Ibmbigdata
Eric Baldeschwieler
 
PDF
"Search, APIs,Capability Management and the Sensis Journey"
Lucidworks (Archived)
 
PPTX
MapR lucidworks joint webinar
Ted Dunning
 
PDF
Non techie journey in social internet age noiselessinnovation
frameworks2go.com
 
PDF
Hadoop's Opportunity to Power Next-Generation Architectures
DataWorks Summit
 
PPTX
Crowd-Sourced Intelligence Built into Search over Hadoop
DataWorks Summit
 
PDF
Introduction to NoSQL and Couchbase
Dipti Borkar
 
A Data Scientist And A Log File Walk Into A Bar...
Paco Nathan
 
Cascading: Enterprise Data Workflows based on Functional Programming
Paco Nathan
 
Cascading meetup #4 @ BlueKai
Paco Nathan
 
Chicago Hadoop Users Group: Enterprise Data Workflows
Paco Nathan
 
NOSQLEU - Graph Databases and Neo4j
Tobias Lindaaker
 
Pattern: an open source project for migrating predictive models onto Apache H...
Paco Nathan
 
Intro to Data Science for Enterprise Big Data
Paco Nathan
 
Search, APIs, capability management and Sensis's journey
ablebagel
 
sones company presentation
sones GmbH
 
Anti-social Databases
William LaForest
 
Good Data: Collaborative Analytics On Demand
zsvoboda
 
Large-Scale Search Discovery Analytics with Hadoop, Mahout, Solr
DataWorks Summit
 
Tech4Africa - Opportunities around Big Data
Steve Watt
 
hadoop @ Ibmbigdata
Eric Baldeschwieler
 
"Search, APIs,Capability Management and the Sensis Journey"
Lucidworks (Archived)
 
MapR lucidworks joint webinar
Ted Dunning
 
Non techie journey in social internet age noiselessinnovation
frameworks2go.com
 
Hadoop's Opportunity to Power Next-Generation Architectures
DataWorks Summit
 
Crowd-Sourced Intelligence Built into Search over Hadoop
DataWorks Summit
 
Introduction to NoSQL and Couchbase
Dipti Borkar
 

More from Paco Nathan (20)

PDF
Human in the loop: a design pattern for managing teams working with ML
Paco Nathan
 
PDF
Human-in-the-loop: a design pattern for managing teams that leverage ML
Paco Nathan
 
PDF
Human-in-a-loop: a design pattern for managing teams which leverage ML
Paco Nathan
 
PDF
Humans in a loop: Jupyter notebooks as a front-end for AI
Paco Nathan
 
PDF
Humans in the loop: AI in open source and industry
Paco Nathan
 
PDF
Computable Content
Paco Nathan
 
PDF
Computable Content: Lessons Learned
Paco Nathan
 
PDF
SF Python Meetup: TextRank in Python
Paco Nathan
 
PDF
Use of standards and related issues in predictive analytics
Paco Nathan
 
PDF
Data Science in 2016: Moving Up
Paco Nathan
 
PDF
Data Science Reinvents Learning?
Paco Nathan
 
PDF
Jupyter for Education: Beyond Gutenberg and Erasmus
Paco Nathan
 
PDF
GalvanizeU Seattle: Eleven Almost-Truisms About Data
Paco Nathan
 
PDF
Microservices, containers, and machine learning
Paco Nathan
 
PDF
GraphX: Graph analytics for insights about developer communities
Paco Nathan
 
PDF
Graph Analytics in Spark
Paco Nathan
 
PDF
Apache Spark and the Emerging Technology Landscape for Big Data
Paco Nathan
 
PDF
QCon São Paulo: Real-Time Analytics with Spark Streaming
Paco Nathan
 
PDF
Strata 2015 Data Preview: Spark, Data Visualization, YARN, and More
Paco Nathan
 
PDF
A New Year in Data Science: ML Unpaused
Paco Nathan
 
Human in the loop: a design pattern for managing teams working with ML
Paco Nathan
 
Human-in-the-loop: a design pattern for managing teams that leverage ML
Paco Nathan
 
Human-in-a-loop: a design pattern for managing teams which leverage ML
Paco Nathan
 
Humans in a loop: Jupyter notebooks as a front-end for AI
Paco Nathan
 
Humans in the loop: AI in open source and industry
Paco Nathan
 
Computable Content
Paco Nathan
 
Computable Content: Lessons Learned
Paco Nathan
 
SF Python Meetup: TextRank in Python
Paco Nathan
 
Use of standards and related issues in predictive analytics
Paco Nathan
 
Data Science in 2016: Moving Up
Paco Nathan
 
Data Science Reinvents Learning?
Paco Nathan
 
Jupyter for Education: Beyond Gutenberg and Erasmus
Paco Nathan
 
GalvanizeU Seattle: Eleven Almost-Truisms About Data
Paco Nathan
 
Microservices, containers, and machine learning
Paco Nathan
 
GraphX: Graph analytics for insights about developer communities
Paco Nathan
 
Graph Analytics in Spark
Paco Nathan
 
Apache Spark and the Emerging Technology Landscape for Big Data
Paco Nathan
 
QCon São Paulo: Real-Time Analytics with Spark Streaming
Paco Nathan
 
Strata 2015 Data Preview: Spark, Data Visualization, YARN, and More
Paco Nathan
 
A New Year in Data Science: ML Unpaused
Paco Nathan
 

Recently uploaded (20)

PDF
Blue Purple Modern Animated Computer Science Presentation.pdf.pdf
Akar16
 
PPTX
Cloud computing and distributed systems.
kkkkkhan
 
PDF
Advanced IT Governance
University of Hertfordshire
 
PDF
Chapter 3 Spatial Domain Image Processing.pdf
Getnet Tigabie Askale -(GM)
 
PPTX
20250228 LYD VKU AI Blended-Learning.pptx
DatVoNgoc
 
PDF
Empathic Computing: Creating Shared Understanding
Mark Billinghurst
 
PDF
Electronic commerce courselecture one. Pdf
OmerMohamed64
 
PDF
Mobile App Security Testing_ A Comprehensive Guide.pdf
flufftailshop
 
PDF
Review of recent advances in non-invasive hemoglobin estimation
IAESIJAI
 
PDF
The Rise and Fall of 3GPP – Time for a Sabbatical?
3G4G
 
PPTX
Detection-First SIEM: Rule Types, Dashboards, and Threat-Informed Strategy
ReZa AdineH
 
PPT
Teaching material agriculture food technology
LiaRayya
 
PDF
Diabetes mellitus diagnosis method based random forest with bat algorithm
IAESIJAI
 
PDF
Modernizing your data center with Dell and AMD
Principled Technologies
 
PDF
solutions_manual_-_materials___processing_in_manufacturing__demargo_.pdf
AbdullahSani29
 
PDF
Advanced Soft Computing BINUS July 2025.pdf
University of Hertfordshire
 
PDF
How UI/UX Design Impacts User Retention in Mobile Apps.pdf
SynapseIndia
 
PPTX
PA Analog/Digital System: The Backbone of Modern Surveillance and Communication
AVTRON Technologies LLC
 
PDF
Advanced methodologies resolving dimensionality complications for autism neur...
IAESIJAI
 
PDF
GDG Cloud Iasi [PUBLIC] Florian Blaga - Unveiling the Evolution of Cybersecur...
athlonica
 
Blue Purple Modern Animated Computer Science Presentation.pdf.pdf
Akar16
 
Cloud computing and distributed systems.
kkkkkhan
 
Advanced IT Governance
University of Hertfordshire
 
Chapter 3 Spatial Domain Image Processing.pdf
Getnet Tigabie Askale -(GM)
 
20250228 LYD VKU AI Blended-Learning.pptx
DatVoNgoc
 
Empathic Computing: Creating Shared Understanding
Mark Billinghurst
 
Electronic commerce courselecture one. Pdf
OmerMohamed64
 
Mobile App Security Testing_ A Comprehensive Guide.pdf
flufftailshop
 
Review of recent advances in non-invasive hemoglobin estimation
IAESIJAI
 
The Rise and Fall of 3GPP – Time for a Sabbatical?
3G4G
 
Detection-First SIEM: Rule Types, Dashboards, and Threat-Informed Strategy
ReZa AdineH
 
Teaching material agriculture food technology
LiaRayya
 
Diabetes mellitus diagnosis method based random forest with bat algorithm
IAESIJAI
 
Modernizing your data center with Dell and AMD
Principled Technologies
 
solutions_manual_-_materials___processing_in_manufacturing__demargo_.pdf
AbdullahSani29
 
Advanced Soft Computing BINUS July 2025.pdf
University of Hertfordshire
 
How UI/UX Design Impacts User Retention in Mobile Apps.pdf
SynapseIndia
 
PA Analog/Digital System: The Backbone of Modern Surveillance and Communication
AVTRON Technologies LLC
 
Advanced methodologies resolving dimensionality complications for autism neur...
IAESIJAI
 
GDG Cloud Iasi [PUBLIC] Florian Blaga - Unveiling the Evolution of Cybersecur...
athlonica
 

Building Enterprise Apps for Big Data with Cascading

  • 1. Building Enterprise Apps for Big Data with Cascading Paco Nathan Document Collection Scrub Tokenize token Concurrent, Inc. M HashJoin Regex Left token GroupBy R Stop Word token List RHS pnathan@concurrentinc.com Count @pacoid Word Count Copyright @2012, Concurrent, Inc.
  • 2. Enterprise Apps for Big Data with Cascading 1. backstory: how we got here 2. build: Data Science teams 3. pattern: common use cases 4. intro: Cascading API 5. tutorial: for the impatient 6. code: sample apps
  • 3. Intro to Cascading Document Collection Scrub Tokenize token M HashJoin Regex Left token GroupBy R Stop Word token List RHS Count Word Count 1. backstory: how we got here
  • 4. inflection point huge Internet successes after 1997 holiday season… 1997 AMZN, EBAY, Inktomi (YHOO Search), then GOOG 1998 consider this metric: annual revenue per customer / amount of data stored which dropped 100x within a few years after 1997 2004 storage and processing costs plummeted, now we must work much smarter to extract ROI from Big Data… our methods must adapt “conventional wisdom” of RDBMS and BI tools became less viable; however, business cadre was still focused on pivot tables and pie charts… which tends toward inertia! MapReduce and the Hadoop open source stack grew directly out of that contention… however, that effort ＋ only solves parts of the puzzle
  • 5. inflection point: consequences Geoffrey Moore (Mohr Davidow Ventures, author of Crossing The Chasm) Hadoop Summit, 2012: “All of Fortune 500 is now on notice over the next 10-year period.” Amazon and Google as exemplars of massive disruption in retail, advertising, etc. data as the major force displacing Global 1000 over the next decade, mostly through apps — verticals, leveraging domain expertise Michael Stonebraker (INGRES, PostgreSQL,Vertica,VoltDB, etc.) XLDB, 2012: “Complex analytics workloads are now displacing SQL as the basis  for Enterprise apps.”
  • 6. primary sources Amazon “Early Amazon: Splitting the website” – Greg Linden glinden.blogspot.com/2006/02/early-amazon-splitting-website.html eBay “The eBay Architecture” – Randy Shoup, Dan Pritchett addsimplicity.com/adding_simplicity_an_engi/2006/11/you_scaled_your.html addsimplicity.com.nyud.net:8080/downloads/eBaySDForum2006-11-29.pdf Inktomi (YHOO Search) “Inktomi’s Wild Ride” – Erik Brewer (0:05:31 ff) youtube.com/watch?v=E91oEn1bnXM Google “The Birth of Google” – John Battelle wired.com/wired/archive/13.08/battelle.html “Underneath the Covers at Google” – Jeff Dean (0:06:54 ff) youtube.com/watch?v=qsan-GQaeyk perspectives.mvdirona.com/2008/06/11/JeffDeanOnGoogleInfrastructure.aspx
  • 7. the world before… BI, SQL, and highly optimized code
  • 8. data innovation: circa 1996 Stakeholder Customers Excel pivot tables PowerPoint slide decks strategy BI Product Analysts requirements SQL Query optimized Engineering code Web App result sets transactions RDBMS
  • 9. the world after… machine learning, leveraging log files
  • 10. data innovation: circa 2001 Stakeholder Product Customers dashboards UX Engineering models servlets recommenders Algorithmic + Web Apps Modeling classifiers Middleware aggregation event SQL Query history result sets customer transactions Logs DW ETL RDBMS
  • 11. the world ahead… what our customers are doing now
  • 12. data innovation: circa 2013 Customers Data Apps business Domain process Workflow Prod Expert dashboard Web Apps, metrics History services Mobile, data etc. s/w science dev Data Planner Scientist social discovery optimized interactions + capacity transactions, Eng endpoints modeling content App Dev Data Access Patterns Hadoop, Log In-Memory etc. Events Data Grid Ops DW Ops batch "real time" Cluster Scheduler introduced existing capability SDLC RDBMS RDBMS
  • 14. statistical thinking Process Variation Data Tools employing a mode of thought which includes both logical and analytical reasoning: evaluating the whole of a problem, as well as its component parts; attempting to assess the effects of changing one or more variables this approach attempts to understand not just problems and solutions, but also the processes involved and their variances particularly valuable in Big Data work when combined with hands-on experience in physics – roughly 50% of my peers come from physics or physical engineering… programmers typically don’t think this way… however, both systems engineers and data scientists must!
  • 15. reference by Leo Breiman Statistical Modeling: The Two Cultures Statistical Science, 2001 bit.ly/eUTh9L
  • 16. Intro to Cascading Document Collection Scrub Tokenize token M HashJoin Regex Left token GroupBy R Stop Word token List RHS Count Word Count 2. build: Data Science teams
  • 17. core values Data Science teams develop actionable insights, building confidence for decisions that work may influence a few decisions worth billions (e.g., M&A) or billions of small decisions (e.g., AdWords) probably somewhere in-between… Wikipedia solving for pattern, at scale. an interdisciplinary pursuit which requires teams, not sole players
  • 18. most valuable skills approximately 80% of the costs for data-related projects get spent on data preparation – mostly on cleaning up data quality issues: ETL, log file analysis, etc. unfortunately, data-related budgets for many companies tend to go into frameworks which can only be used after clean up most valuable skills: ‣ learn to use programmable tools that prepare data ‣ learn to generate compelling data visualizations ‣ learn to estimate the confidence for reported results ‣ learn to automate work, making analysis repeatable D3 the rest of the skills – modeling, algorithms, etc. – those are secondary
  • 19. social caveats “This data cannot be correct!” may be an early warning about an organization itself much depends on how the people whom you work alongside tend to arrive at decisions: ‣ probably good: Induction, Abduction, Circumscription ‣ probably poor: Deduction, Speculation, Justification in general, one good data visualization puts many ongoing verbal arguments to rest however, let domain experts handle “data storytelling”, not data scientists xkcd
  • 20. the science in data science? edoMpUsserD:IUN tcudorP ylppA lenaP yrotnevnI tneilC tcudorP evomeR lenaP yrotnevnI tneilC in a nutshell, what we do… edoMmooRyM:IUN edoMmooRcilbuP:IUN ydduB ddA nigoL etisbeW vd edoMsdneirF:IUN edoMtahC:IUN egasseM a evaeL G1 :gniniamer ecaps sserddA dekcilCeliforPyM:IUN edoMstiderCyuB:IUN tohspanS a ekaT egapemoH nwO tisiV elbbuB a epyT taeS egnahC wodniW D3 nepO dneirF ddA revO tcudorP pilF lenaP yrotnevnI tneilC lenaP tidE ‣ estimate probability woN tahC teP yalP teP deeF 2 petS egaP traC esahcruP edaM remotsuC M215 :gniniamer ecaps sserddA gnihtolC no tuP bew :metI na yuB edoMeivoM:IUN ytinummoc ,tneilc :detratS weiV eivoM teP weN etaerC detrats etius tset :tseTytivitcennoC emag pazyeh dehcnuaL eciov mooRcilbuP tahC egasseM yadhtriB edoMlairotuT:IUN ybbol semag dehcnuaL noitartsigeR euqinU ‣ calculate analytic variance edoMpUsserD:IUN tcudorP ylppA lenaP yrotnevnI tneilC tcudorP evomeR lenaP yrotnevnI tneilC edoMmooRyM:IUN edoMmooRcilbuP:IUN ydduB ddA nigoL etisbeW vd edoMsdneirF:IUN edoMtahC:IUN egasseM a evaeL G1 :gniniamer ecaps sserddA dekcilCeliforPyM:IUN edoMstiderCyuB:IUN tohspanS a ekaT egapemoH nwO tisiV elbbuB a epyT taeS egnahC dneirF ddA revO tcudorP pilF lenaP yrotnevnI tneilC lenaP tidE woN tahC teP yalP teP deeF 2 petS egaP traC esahcruP edaM remotsuC M215 :gniniamer ecaps sserddA gnihtolC no tuP bew :metI na yuB edoMeivoM:IUN ytinummoc ,tneilc :detratS weiV eivoM teP weN etaerC detrats etius tset :tseTytivitcennoC emag pazyeh dehcnuaL eciov mooRcilbuP tahC egasseM yadhtriB edoMlairotuT:IUN ybbol semag dehcnuaL noitartsigeR euqinU wodniW D3 nepO ‣ manipulate order complexity ‣ make use of learning theory + collab with DevOps, Stakeholders + reduce our work to cron entries
  • 21. synthesis of the above MapReduce is Good Enough? Jimmy Lin, U Maryland + Twitter arxiv.org/pdf/1209.2191v1.pdf A Few Useful Things to Know about Machine Learning Pedro Domingos, U Washington homes.cs.washington.edu/~pedrod/papers/cacm12.pdf
  • 22. team process = needs help people ask the discovery right questions allow automation to place modeling informed bets deliver products at integration scale to customers build smarts into apps product features Gephi keep infrastructure systems running, cost-effective
  • 23. team composition = roles Domain Expert business process, stakeholder data science Data data prep, discovery, Scientist modeling, etc. Document Collection Scrub Tokenize token M HashJoin Regex Left token GroupBy R Stop Word token List RHS App Dev software engineering, Count automation Word Count Ops systems engineering, access introduced capability
  • 24. matrix = needs × roles nn o overy very elliing e ng ratiio rat o apps apps tem tem ss diisc d sc mod mod nteg ii nteg sys sys stakeholder scientist developer ops
  • 25. matrix: example team nn o overy very elliing e ng ratiio rat o apps apps tem tem ss diisc d sc mod mod nteg ii nteg sys sys stakeholder scientist developer ops summary: this team seems heavy on systems, may need more overlap between modeling and integration, particularly among team leads
  • 26. Q: Can I simply hire one rockstar data scientist to cover all this work?
  • 27. A: No, interdisciplinary work requires teams. A: Hire leads who speak the lingo of each domain. A: Hire people who cover 2+ roles, when possible.
  • 28. reference by DJ Patil Data Jujitsu O’Reilly, 2012 amazon.com/dp/B008HMN5BE Building Data Science Teams O’Reilly, 2011 amazon.com/dp/B005O4U3ZE
  • 29. Intro to Cascading Document Collection Scrub Tokenize token M HashJoin Regex Left token GroupBy R Stop Word token List RHS Count Word Count 3. pattern: common use cases
  • 30. CAP theorem purpose: theoretical limits for data access patterns essence: ‣ consistency ‣ availability ‣ partition tolerance best case scenario: you may pick two … or spend billions struggling to obtain all three at scale (GOOG) translated: cost of doing business www.cs.berkeley.edu/~brewer/cs262b-2004/PODC-keynote.pdf julianbrowne.com/article/viewer/brewers-cap-theorem
  • 31. data access patterns design patterns: originated in consensus negotiation for architecture, later used in software engineering consider the corollaries in large-scale data work… essence: select data frameworks based on your data access patterns in other words, decouple use cases based on needs – avoid the “one size fits all” (OSFA) anti-pattern let’s review some examples…
  • 32. access → frameworks → forfeits financial transactions general ledger in RDBMS CAx ad-hoc queries RDS (hosted MySQL) CAx reporting, dashboards like Pentaho CAx log rotation/persistence like Riak xxP search indexes like Lucene/Solr xAP static content, archives S3 (durable storage) xAP customer facts like Redis, Membase xAP distributed counters, locks, sets like Redis x A P* data objects CRUD key/value – like, NoSQL on MySQL CxP authoritative metadata like Zookeeper CxP data prep, modeling at scale like Hadoop/Cascading + R CxP graph analysis like Hadoop + Redis + Gephi CxP data marts like Hadoop/HBase CxP
  • 33. access → frameworks → forfeits financial transactions general ledger in RDBMS CAx ad-hoc queries RDS (hosted MySQL) CAx reporting, dashboards like Pentaho CAx log rotation/persistence like Riak xxP search indexes like Lucene/Solr xAP static content, archives S3 (durable storage) xAP customer facts like Redis, Membase xAP distributed counters, locks, sets like Redis x A P* data objects CRUD key/value – like, NoSQL on MySQL CxP authoritative metadata like Zookeeper CxP data prep, modeling at scale like Hadoop/Cascading + R CxP graph analysis like Hadoop + Redis + Gephi CxP data marts like Hadoop/HBase CxP
  • 34. and, since “One Size Fits All”… doesn’t
  • 35. a selection of great tools… reporting: visualization: Graphite, PowerPivot, ggplot2, D3, Gephi analytics/modeling: Pentaho, Jaspersoft, SAS R, Weka, Matlab, PMML, GLPK text: LDA, WordNet, OpenNLP, Mallet, Bixo, NLTK apps: Cascading, Scalding, Cascalog, R markdown, SWF scale-out: Scalr, RightScale, CycleComputing, vFabric, Beanstalk graph: column: Gremlin, Vertica, GraphLab, HBase, key/val: index: relational: Neo4J Drill, Redis, Lucene/Solr, usual suspects Dynamo Membase, ElasticSearch MySQL imdg: Spark, Storm, hadoop: EMR, HW, MapR, machine data: Gigaspaces EMC, Azure, Compute Splunk, collectd, durable storage: Nagios S3, ASV, GCS, Riak, Couch
  • 36. common use cases app patterns
  • 37. use case: marketing funnel • must optimize a very large ad spend • different vendors report different metrics Wikipedia • seasonal variation distorts performance • some campaigns are much smaller than others • hard to predict ROI for incremental spend approach: • log aggregation, followed with cohort analysis • bayesian point estimates compare different-sized ad tests • customer lifetime value quantifies ROI of new leads • time series analysis normalizes for seasonal variation • geolocation adjusts for regional cost/benefit • linear programming models estimate elasticity of demand
  • 38. use case: ecommerce fraud • sparse data means lots of missing values stat.berkeley.edu • “needle in a haystack” lack of training cases • answers are available in large-scale batch, results are needed in real-time event processing • not just one pattern to detect – many, ever-changing approach: • random forest (RF) classifiers predict likely fraud • subsampled data to re-balance training sets • impute missing values based on density functions • train on massive log files, run on in-memory grid • adjust metrics to minimize customer support costs • detect novelty – report anomalies via notifications
  • 39. use case: customer segmentation • many millions of customers, hard to determine which features resonate Mathworks • multi-modal distributions get obscured by the practice of calculating an “average” • not much is known about individual customers approach: • connected components for sessionization, determining uniques from logs • estimates for age, gender, income, geo, etc. • clustering algorithms to group into market segments • social graph infers “unknown” relationships • covariance/heat maps visualizes segments vs. feature sets
  • 40. use case: monetizing content • need to suggest relevant content which would Digital Humanities otherwise get buried in the back catalog • big disconnect between inventory and limited performance ad market • enormous amounts of text, hard to categorize approach: • text analytics glean key phrases from documents • hierarchical clustering of char frequencies detects lang • latent dirichlet allocation (LDA) reduces dimension to topic models • recommenders suggest similar topics to customers • collaborative filters connect known users with less known
  • 41. Intro to Cascading Document Collection Scrub Tokenize token M HashJoin Regex Left token GroupBy R Stop Word token List RHS Count Word Count 4. intro: Cascading API
  • 42. Cascading API: purpose ‣ simplify data processing development and deployment ‣ improve application developer productivity ‣ enable data processing application manageability
  • 43. Cascading API: a few facts Java open source project (ASL 2) using Git, Gradle, Maven, JUnit, etc. in production (~5 yrs) at hundreds of enterprise Hadoop deployments: Finance, Health Care, Transportation, other verticals studies published about large use cases: Twitter, Etsy, Airbnb, Square, Climate Corporation, FlightCaster, Williams-Sonoma partnerships and distribution with SpringSource, Amazon AWS, Microsoft Azure, Hortonworks, MapR, EMC several open source projects built atop, managed by Twitter, Etsy, etc., which provide substantial Machine Learning libraries DSLs available in Scala, Clojure, Python (Jython), Ruby (JRuby), Groovy data “taps” integrate popular data frameworks via JDBC, Memcached, HBase, plus serialization in Apache Thrift, Avro, Kyro, etc. entire app compiles into a single JAR: fully connected for compiler optimization, exception handling, debugging, config, scheduling, etc.
  • 44. Cascading API: a few quotes “Cascading gives Java developers the ability to build Big Data applications on Hadoop using their existing skillset … Management can really go out and build a team around folks that are already very experienced with Java. Switching over to this is really a very short exercise.” CIO, Thor Olavsrud, 2012-06-06 cio.com/article/707782/Ease_Big_Data_Hiring_Pain_With_Cascading “Masks the complexity of MapReduce, simplifies the programming, and speeds you on your journey toward actionable analytics … A vast improvement over native MapReduce functions or Pig UDFs.” 2012 BOSSIE Awards, James Borck, 2012-09-18 infoworld.com/slideshow/65089 “Company’s promise to application developers is an opportunity to build and test applications on their desktops in the language of choice with familiar constructs and reusable components” Dr. Dobb’s, Adrian Bridgwater, 2012-06-08 drdobbs.com/jvm/where-does-big-data-go-to-get-data-inten/240001759
  • 45. data+code “political spectrum” “Notes from the Mystery Machine Bus” by Steve Yegge, Google goo.gl/SeRZa “conservative” “liberal” (mostly) Enterprise (mostly) Start-Up risk management customer experiments assurance flexibility well-defined schema schema follows code explicit configuration convention type-checking compiler interpreted scripts wants no surprises wants no impediments Java, Scala, Clojure, etc. PHP, Ruby, Python, etc. Cascading, Scalding, Cascalog, etc. Hive, Pig, Hadoop Streaming, etc.
  • 46. Cascading API: adoption As Enterprise apps move into Hadoop and related BigData frameworks, risk profiles shift toward more conservative programming practices Cascading provides a popular API for defining and managing Enterprise data workflows
  • 47. enterprise data workflows Tuples, Pipelines, Endpoints, Operations, Joins, Assertions, Traps, etc. …in other words, “plumbing” Document Collection Scrub Tokenize token M HashJoin Regex Left token GroupBy R Stop Word token List RHS Count Word Count
  • 48. data workflows: team ‣ Business Stakeholder POV: business process management for workflow orchestration (think BPM/BPEL) ‣ Systems Integrator POV: system integration of heterogenous data sources and compute platforms ‣ Data Scientist POV: a directed, acyclic graph (DAG) on which we can apply Amdahl's Law, etc. ‣ Data Architect POV: a physical plan for large-scale data flow management ‣ Software Architect POV: a pattern language, similar to plumbing or circuit design Document Collection ‣ App Developer POV: M Tokenize Scrub token API bindings for Java, Scala, Clojure, Jython, JRuby, etc. Stop Word List HashJoin Left RHS Regex token GroupBy token R Count ‣ Systems Engineer POV: Word Count a JAR file, has passed CI, available in a Maven repo
  • 49. data workflows: layers business domain expertise, business trade-offs, process operating parameters, market position, etc. API Java, Scala, Clojure, Jython, JRuby, Groovy, etc. language …envision whatever runs in a JVM optimize / schedule major changes in technology now Document Collection Scrub Tokenize token physical M HashJoin Regex Left token GroupBy R plan Stop Word token List RHS Count Word Count compute Apache Hadoop, in-memory local mode “assembler” code substrate …envision GPUs, streaming, etc. machine data Splunk, Nagios, Collectd, New Relic, etc.
  • 50. data workflows: SQL Relational SQL parser logical plan, optimized based on stats physical plan query history, table stats b-trees, etc. ERD table schema catalog
  • 51. data workflows: SQL vs. JVM Relational Cascading + Driven SQL parser SQL-92 compliant parser (in progress) logical plan, TODO: logical plan, optimized based on stats optimized based on stats physical plan API “plumbing” query history, app history, table stats tuple stats b-trees, etc. distributed compute substrate: Hadoop, in-memory, etc. ERD flow diagram table schema tuple schema catalog endpoint usage DB
  • 52. Intro to Cascading Document Collection Scrub Tokenize token M HashJoin Regex Left token GroupBy R Stop Word token List RHS Count Word Count 5. tutorial: for the impatient
  • 53. “Cascading for the Impatient” cascading.org/category/impatient/ ‣ a series of introductory tutorials and code samples ‣ 1:1 code comparisons in Scalding, Cascalog, Pig, Hive Document Collection Scrub Tokenize token M HashJoin Regex Left token GroupBy R Stop Word token List RHS Count Word Count
  • 54. 1: copy public class   Main   {   public static void   main( String[] args )     {     String inPath = args[ 0 ];     String outPath = args[ 1 ]; Source     Properties props = new Properties();     AppProps.setApplicationJarClass( props, Main.class );     HadoopFlowConnector flowConnector = new HadoopFlowConnector( props );     // create the source tap     Tap inTap = new Hfs( new TextDelimited( true, "t" ), inPath ); M     // create the sink tap     Tap outTap = new Hfs( new TextDelimited( true, "t" ), outPath ); Sink     // specify a pipe to connect the taps     Pipe copyPipe = new Pipe( "copy" );     // connect the taps, pipes, etc., into a flow     FlowDef flowDef = FlowDef.flowDef().setName( "copy" )      .addSource( copyPipe, inTap )      .addTailSink( copyPipe, outTap );     // run the flow     flowConnector.connect( flowDef ).complete(); 1 mapper     }   } 0 reducers 10 lines code
  • 55. wait! ten lines of code for a file copy… seems like a lot.
  • 56. same JAR, any scale… MegaCorp Enterprise IT: Pb’s data 1000+ node private cluster EVP calls you when app fails runtime: days+ Production Cluster: Tb’s data EMR w/ 50 HPC Instances Ops monitors results runtime: hours – days Staging Cluster: Gb’s data EMR + 4 Spot Instances CI shows red or green lights runtime: minutes – hours Your Laptop: Mb’s data Hadoop standalone mode passes unit tests, or not runtime: seconds – minutes
  • 57. 2: word count Document Collection Tokenize GroupBy M token Count R Word Count 1 mapper 1 reducer 18 lines code gist.github.com/3900702
  • 58. Cascading / Java Document Collection M Tokenize GroupBy token Count String docPath = args[ 0 ]; R Word String wcPath = args[ 1 ]; Count Properties properties = new Properties(); AppProps.setApplicationJarClass( properties, Main.class ); HadoopFlowConnector flowConnector = new HadoopFlowConnector( properties ); // create source and sink taps Tap docTap = new Hfs( new TextDelimited( true, "t" ), docPath ); Tap wcTap = new Hfs( new TextDelimited( true, "t" ), wcPath ); // specify a regex to split "document" text lines into token stream Fields token = new Fields( "token" ); Fields text = new Fields( "text" ); RegexSplitGenerator splitter = new RegexSplitGenerator( token, "[ [](),.]" ); // only returns "token" Pipe docPipe = new Each( "token", text, splitter, Fields.RESULTS ); // determine the word counts Pipe wcPipe = new Pipe( "wc", docPipe ); wcPipe = new GroupBy( wcPipe, token ); wcPipe = new Every( wcPipe, Fields.ALL, new Count(), Fields.ALL ); // connect the taps, pipes, etc., into a flow FlowDef flowDef = FlowDef.flowDef().setName( "wc" ) .addSource( docPipe, docTap )  .addTailSink( wcPipe, wcTap ); // write a DOT file and run the flow Flow wcFlow = flowConnector.connect( flowDef ); wcFlow.writeDOT( "dot/wc.dot" ); wcFlow.complete();
  • 59. Scalding / Scala Document Collection M Tokenize GroupBy token Count R Word Count // Sujit Pal // sujitpal.blogspot.com/2012/08/scalding-for-impatient.html package com.mycompany.impatient import com.twitter.scalding._ class Part2(args : Args) extends Job(args) {   val input = Tsv(args("input"), ('docId, 'text))   val output = Tsv(args("output"))   input.read.     flatMap('text -> 'word) { text : String => text.split("""s+""") }.     groupBy('word) { group => group.size }.     write(output) }
  • 60. Cascalog / Clojure Document Collection M Tokenize GroupBy token Count R Word Count ; Paul Lam ; github.com/Quantisan/Impatient (ns impatient.core   (:use [cascalog.api]         [cascalog.more-taps :only (hfs-delimited)])   (:require [clojure.string :as s]             [cascalog.ops :as c])   (:gen-class)) (defmapcatop split [line]   "reads in a line of string and splits it by regex"   (s/split line #"[[](),.)s]+")) (defn -main [in out & args]   (?<- (hfs-delimited out)        [?word ?count]        ((hfs-delimited in :skip-header? true) _ ?line)        (split ?line :> ?word)        (c/count ?count)))
  • 61. Hive Document Collection M Tokenize GroupBy token Count R Word Count -- Steve Severance -- stackoverflow.com/questions/10039949/word-count-program-in-hive CREATE TABLE input (line STRING); LOAD DATA LOCAL INPATH 'input.tsv' OVERWRITE INTO TABLE input; SELECT  word, COUNT(*) FROM input  LATERAL VIEW explode(split(text, ' ')) lTable AS word GROUP BY word ;
  • 62. Pig Document Collection M Tokenize GroupBy token Count R Word Count -- kudos to Dmitriy Ryaboy docPipe = LOAD '$docPath' USING PigStorage('t', 'tagsource') AS (doc_id, text); docPipe = FILTER docPipe BY doc_id != 'doc_id'; -- specify regex to split "document" text lines into token stream tokenPipe = FOREACH docPipe GENERATE doc_id, FLATTEN(TOKENIZE(text, ' [](),.')) AS token; tokenPipe = FILTER tokenPipe BY token MATCHES 'w.*'; -- determine the word counts tokenGroups = GROUP tokenPipe BY token; wcPipe = FOREACH tokenGroups GENERATE group AS token, COUNT(tokenPipe) AS count; -- output STORE wcPipe INTO '$wcPath' USING PigStorage('t', 'tagsource'); EXPLAIN -out dot/wc_pig.dot -dot wcPipe;
  • 63. 3: wc + scrub Document Collection Scrub GroupBy Tokenize token token Count M R Word Count 1 mapper 1 reducer 22+10 lines code
  • 64. 4: wc + scrub + stop words Document Collection Scrub Tokenize token M HashJoin Regex Left token GroupBy R Stop Word token List RHS Count 1 mapper Word 1 reducer Count 28+10 lines code
  • 65. 5: tf-idf Unique Insert SumBy D doc_id 1 doc_id Document Collection M R M R M RHS Scrub Tokenize token HashJoin M RHS token HashJoin Regex Unique GroupBy DF Left token token token ExprFunc Count CoGroup Stop Word tf-idf List RHS M R M R M R TF-IDF M GroupBy TF doc_id, token Count GroupBy Count token M R M R Word R M R Count 11 mappers 9 reducers 65+10 lines code
  • 66. 6: tf-idf + tdd Unique Insert SumBy D doc_id 1 doc_id Document Collection RHS M R M R M Assert Scrub Tokenize token HashJoin Checkpoint M M RHS token HashJoin Regex Unique GroupBy DF Left token token token Count ExprFunc CoGroup tf-idf Stop Word List RHS M R M R M R TF-IDF M GroupBy TF doc_id, Failure token Count Traps GroupBy Count token M R M R Word Count R M R 12 mappers 9 reducers 76+14 lines code
  • 67. deployed on AWS… elastic-mapreduce --create --name "TF-IDF" --jar s3n://temp.cascading.org/impatient/part6.jar --arg s3n://temp.cascading.org/impatient/rain.txt --arg s3n://temp.cascading.org/impatient/out/wc --arg s3n://temp.cascading.org/impatient/en.stop --arg s3n://temp.cascading.org/impatient/out/tfidf --arg s3n://temp.cascading.org/impatient/out/trap --arg s3n://temp.cascading.org/impatient/out/check aws.amazon.com/elasticmapreduce/
  • 68. results? doc_id tf-idf doc02 0.9163 token air doc05 0.9163 australia doc05 0.9163 broken doc04 0.9163 california's doc04 0.9163 cause doc02 0.9163 cloudcover doc04 0.9163 death doc04 0.9163 deserts doc03 0.9163 downwind doc_id text … doc01 A rain shadow is a dry area on the lee back side of a mountainous area. doc02 0.9163 sinking doc02 This sinking, dry air produces a rain shadow, or area in the lee of a mountain doc04 0.9163 such with less rain and cloudcover. doc04 0.9163 valley doc03 A rain shadow is an area of dry land that lies on the leeward (or downwind) doc05 0.9163 women side of a mountain. doc03 0.5108 land doc04 This is known as the rain shadow effect and is the primary cause of leeward doc05 0.5108 land deserts of mountain ranges, such as California's Death Valley. doc01 0.5108 lee doc05 Two Women. Secrets. A Broken Land. [DVD Australia] doc02 0.5108 lee zoink null doc03 0.5108 leeward doc04 0.5108 leeward doc01 0.4463 area doc02 0.2231 area doc03 0.2231 area doc01 0.2231 dry doc02 0.2231 dry doc03 0.2231 dry doc02 0.2231 mountain Unique Insert SumBy D doc_id 1 doc_id Document Collection RHS M R M R M doc03 0.2231 mountain Assert Scrub Tokenize token HashJoin Checkpoint M M RHS token HashJoin Regex Unique GroupBy DF Left token doc04 0.2231 mountain token token Count ExprFunc CoGroup tf-idf Stop Word List RHS M R M R M R TF-IDF GroupBy M doc01 0.0000 rain TF doc_id, Failure token Count Traps GroupBy Count token doc02 0.0000 rain M R M R Word Count R M R doc03 0.0000 rain doc04 0.0000 rain doc01 0.0000 shadow doc02 0.0000 shadow doc03 0.0000 shadow doc04 0.0000 shadow
  • 69. comparisons? compare similar code in Scalding (Scala) and Cascalog (Clojure): sujitpal.blogspot.com/2012/08/scalding-for-impatient.html based on: github.com/twitter/scalding/wiki github.com/Quantisan/Impatient based on: github.com/nathanmarz/cascalog/wiki
  • 70. Intro to Cascading Document Collection Scrub Tokenize token M HashJoin Regex Left token GroupBy R Stop Word token List RHS Count Word Count 6. code: sample apps
  • 71. Social Recommender filter Twitter stop words tweets calculate QA similiarity threshold min, max Neo4j LDA Redis github.com/Cascading/SampleRecommender ‣ social recommender based on Twitter: suggest users who tweet about similar stocks ‣ instead of a cross-product (potential bottleneck) this runs in parallel on Hadoop ‣ uses a stop word list to remove common words, offensive phrases, etc. ‣ one tap measures token frequency: for QA, adjust stop words, improve filter, etc. ‣ adapted in Spring by Costin Leau
  • 72. SocRec: architecture Twitter filter low-freq firehose source stop words tweets batch updates ( uid, tweet, t ) checkpoint: tokenized tweets calculate checkpoint: analysis + QA similiarity token frequency curation checkpoint: similarity similar users thresholds threshold min, max sink sink sink Neo4j: social Redis graph LDA: topic results (uid: uidx, rank) trending
  • 73. SocRec: results uid recommend weight carbonfiberxrm ClosingBellNews 0.1459 carbonfiberxrm DJFunkyGrrL 0.0870 ClosingBellNews DJFunkyGrrL 0.1491 CloudStocks DJFunkyGrrL 0.1206 ElmoreNicole DJFunkyGrrL 0.1798 EsNeey alexiolo_ 0.8603 ...
  • 74. City of Palo Alto open data Regex Regex tree Scrub filter parser species M HashJoin Left Geohash CoPA GIS exprot Tree Metadata M RHS RHS tree Regex Checkpoint road Regex Regex tsv parser tsv filter Tree Filter GroupBy Checkpoint parser CoGroup Distance tree_dist tree_name shade M R M R M RHS M HashJoin Estimate Road Left Albedo Segments Geohash CoGroup Road Metadata GPS Failure RHS M logs Traps R road Geohash M Regex park filter reco M park github.com/Cascading/CoPA/wiki ‣ GIS export for parks, roads, trees (unstructured / open data) ‣ log files of personalized/frequented locations in Palo Alto via iPhone GPS tracks ‣ curated metadata, used to enrich the dataset ‣ could extend via mash-up with many available public data APIs Enterprise-scale app: road albedo + tree species metadata + geospatial indexing “Find a shady spot on a summer day to walk near downtown and take a call…”
  • 76. CoPA: results 0.12 Estimated Tree Height (meters) 0.10 0.08 count 0 density 100 0.06 200 300 0.04 0.02 0.00 0 10 20 30 40 50 avg_height ‣ addr: 115 HAWTHORNE AVE ‣ lat/lng: 37.446, -122.168 ‣ geohash: 9q9jh0 ‣ tree: 413 site 2 ‣ species: Liquidambar styraciflua ‣ avg height 23 m ‣ road albedo: 0.12 ‣ distance: 10 m ‣ a short walk from my train stop ✔
  • 77. drill-down blog, code/wiki/gists, jars, list, DevOps products: cascading.org/ github.org/Cascading/ conjars.org/ goo.gl/KQtUL concurrentinc.com/ pnathan@concurrentinc.com @pacoid

Editor's Notes