SlideShare a Scribd company logo
www.africastalking.com 1The Business of APIs
Building Analytics and
Data Science at a
Startup
Kathleen Siminyu
Head of Data Science
● Mathematics and Computer Science
background
● Self-taught data scientist
● Co-organiser of Nairobi WiMLDS (We
turn 2 on Dec. 8th!)
● Organising Committee Deep Learning
Indaba
● Head of Data Science at Africa’s
Talking
www.africastalking.com 3The Business of APIs
Africa’s Talking is a Pan-African, Platform as a Service (PaaS) company based in Nairobi, Kenya.
Our world-class infrastructure is created for developers and relied upon by businesses.
We avail the complexity of SMS, VOICE, USSD, AIRTIME & PAYMENTS via simple API calls and at a fraction of the cost.
ABOUT AFRICA’S TALKING
One Continent. Unified platform. Unimaginable possibilities.
www.africastalking.com 4The Business of APIs
By The Numbers
Active Countries
Planned Countries 2018
Nigeria
Ethiopia
Côte d’Ivoire Kenya
Botswana
Tanzania
Uganda
Rwanda
Malawi
Zimbabwe
Zambia
DR Congo
Sierra Leone
Senegal
Gambia
Ghana
South Africa
Mauritius
11 Countries
20K+ Developers
30M+ API Calls Daily
5K+ Businesses
www.africastalking.com 5The Business of APIs
Airtime
Pinless airtime top-up
from 1 wallet across 4
markets
SMS
2-Way Bulk and
Premium SMS APIs
across 6 Markets
Payments
One wallet for all your
mobile, bank and card
payments across 5
Markets
USSD
Unified Push/Pull USSD API
across 6 Markets
Build intelligent cloud-based Voice applications
VOICE
OUR APIs
www.africastalking.com 6The Business of APIs
Echo Mobile
High impact mobile
surveys via SMS, Voice,
Airtime and USSD
iShamba
Using Voice & SMS to
connect agriculturalists
to experts in the field
M-Kopa
Customer engagement
through SMS to provide
affordable solar to off-grid
homes
CIC Bima
Providing life
insurance through
airtime
Naivas
Bulk SMS for
promotional content &
customer engagement
mShule
Quality and personalized
learning through SMS,
USSD and voice
USE CASES
Over 5,000 business are currently consuming our services. Here’s how:
www.africastalking.com 7The Business of APIs
Session Overview
• What We Did
• Lessons Learnt
www.africastalking.com 8The Business of APIs
What We Did
Technical assessment of our data situation - taking stock
• Our core infrastructure is written in Scala
• Analytics service powered by Redis, Cassandra and Akka Actors
- mainly to service client dashboards
• MySQL databases with client info
• Different softwares used by finance, client relations,
• Applications logs with everything else that didn’t make it into a
DB
www.africastalking.com 9The Business of APIs
What We Did
Explore opportunities and use cases for analytics and data science
• Clustering of clients and markets
• Time series predictions to predict traffic on our infrastructure
• Anomaly detection, especially for our payment products
• Building developer profiles using features like programming
languages, errors, sophistication
www.africastalking.com 10The Business of APIs
What We Did
Explore architectures and tools that might suit our needs
• Settled on a lambda architecture ---> allows us to run both
batch and streaming analytics jobs
• Our stack
• Apache Kafka
• Apache Spark
• HDFS from Apache Hadoop collection of softwares
• Apache HBase
• Scala - programming language of choice
• Docker
• Apache Mesos
• Kubernetes
• Apache Avro
• Apache parquet
www.africastalking.com 11The Business of APIs
What We Did
Build the team in-house
• Hiring proved a challenge, almost no one in the ecosystem had
experience with these tools
• Changed tactics and convinced several engineers to slowly start
porting their skills to become data engineers
www.africastalking.com 12The Business of APIs
Lessons Learnt
• Have a dedicated team solely focused on data
• Distributed systems can be a hairy situation esp security and
logging
• It’s important to manage expectations from the rest of the
organisation
• Buy in from the whole organisation is important to avoid
duplication of efforts
• Hold off hiring until you know what you are building and what
kind of talent you need
www.africastalking.com 13The Business of APIs
THANK YOU

More Related Content

PDF
apidays LIVE New York 2021 - How to productify your data with a GraphQL APIs ...
PDF
Automating Business Processes with APIs
PPTX
SAP Integration for non developers Teched 2017
PPTX
Mule iON - OSS ESB to iPaaS
PPTX
Digital Transformation: Connected API Ecosystems
PPTX
SAP HANA IoT Platform
PPTX
APIdays Paris 2018 - Europ Assistance, the transformation through Public APIs...
PPTX
Chapter deck january 2015
apidays LIVE New York 2021 - How to productify your data with a GraphQL APIs ...
Automating Business Processes with APIs
SAP Integration for non developers Teched 2017
Mule iON - OSS ESB to iPaaS
Digital Transformation: Connected API Ecosystems
SAP HANA IoT Platform
APIdays Paris 2018 - Europ Assistance, the transformation through Public APIs...
Chapter deck january 2015

What's hot (20)

PPTX
Company Profile - Saaragh Technologies
PDF
How Websites go Serverless - WebSummit Lisbon 2018
PDF
"DeepRacer, not just a fancy racecar. ALSO a fancy racecar" - Abby Dryer AWS...
PPTX
Sarine's Big Data Journey by Rostislav Aaronov
PPTX
ReST API's for IBMi Made Simple LXR v1.6.9
PDF
Alexa, Chatbots & Co: Getting ready for the Post-Web World with a Headless Ap...
PPT
Cloudera's Original Pitch Deck from 2008
PDF
[WSO2Con Asia 2018] Qantas APIs Take Flight
PPTX
Big Data Ecosystem- Impetus Technologies
PDF
Cloud expo june 2013: Building a Real Time Analytics Platform on Big Data in ...
PDF
apidays LIVE Australia 2021 - Composable data for the composable enterprise b...
PDF
Roadshow September 2018
PPTX
Session01 introduction to_cloud_by_kalyan_mahalingam
PPTX
Saaragh Technologies Profile 2016
PDF
Apache Kafka in the Automotive Industry (Connected Vehicles, Manufacturing 4....
PDF
SnapLogic Elastic Integration Platform as a Service (iPaaS)
PDF
Best Practices for Streaming IoT Data with MQTT and Apache Kafka
PPTX
London Adapt or Die: Opening Keynote with Chet Kapoor
PDF
Arina Programming
PDF
INJECT MACHINE LEARNING AND OTHER CLOUD SERVICES IN YOUR EXISTING APPLICATIONS
Company Profile - Saaragh Technologies
How Websites go Serverless - WebSummit Lisbon 2018
"DeepRacer, not just a fancy racecar. ALSO a fancy racecar" - Abby Dryer AWS...
Sarine's Big Data Journey by Rostislav Aaronov
ReST API's for IBMi Made Simple LXR v1.6.9
Alexa, Chatbots & Co: Getting ready for the Post-Web World with a Headless Ap...
Cloudera's Original Pitch Deck from 2008
[WSO2Con Asia 2018] Qantas APIs Take Flight
Big Data Ecosystem- Impetus Technologies
Cloud expo june 2013: Building a Real Time Analytics Platform on Big Data in ...
apidays LIVE Australia 2021 - Composable data for the composable enterprise b...
Roadshow September 2018
Session01 introduction to_cloud_by_kalyan_mahalingam
Saaragh Technologies Profile 2016
Apache Kafka in the Automotive Industry (Connected Vehicles, Manufacturing 4....
SnapLogic Elastic Integration Platform as a Service (iPaaS)
Best Practices for Streaming IoT Data with MQTT and Apache Kafka
London Adapt or Die: Opening Keynote with Chet Kapoor
Arina Programming
INJECT MACHINE LEARNING AND OTHER CLOUD SERVICES IN YOUR EXISTING APPLICATIONS
Ad

Similar to Building Analytics and Data Science at A Start-Up, by Kathleen Siminyu, Head of Data Science at Africa's Talking (20)

PDF
PWA 101, what you need to know about ShopwarePWA
DOC
Varun-CV-J
PDF
Market_Cloud_AI_Capabilities_POC Demo.pdf
PDF
APEX Alpe Adria Mike Hichwa Keynote April 11th 2019- Zagreb
PPTX
Deploy Apache Spark™ on Rackspace OnMetal™ for Cloud Big Data Platform
PDF
Webinar: iPaaS in the Enterprise - What to Look for in a Cloud Integration Pl...
PDF
0626 2014 01_toronto-smac meetup_io_t
PPTX
Bluegrass digital creds 2015
PDF
Introduction to NEW SAP - Accenture Technology Meetup
PDF
AWS Summit Berlin 2013 - Next-Generation Websites on Demand – with Infopark C...
PDF
Kafka: Journey from Just Another Software to Being a Critical Part of PayPal ...
PPTX
DataStax on Azure: Deploying an industry-leading data platform for cloud apps...
PPTX
About Tekmonks
PPTX
Agoda open stack in a large scale deployment
PDF
SAP LEONARDO SAP LEONARDO the digital digital innovation innovation innovatio...
PPT
Power point updated
PDF
Application Services - CCS Technologies (P) Ltd.
PDF
meet4psa
PDF
Can Your Mobile Infrastructure Survive 1 Million Concurrent Users?
PDF
Digital Reinvention by NRB
PWA 101, what you need to know about ShopwarePWA
Varun-CV-J
Market_Cloud_AI_Capabilities_POC Demo.pdf
APEX Alpe Adria Mike Hichwa Keynote April 11th 2019- Zagreb
Deploy Apache Spark™ on Rackspace OnMetal™ for Cloud Big Data Platform
Webinar: iPaaS in the Enterprise - What to Look for in a Cloud Integration Pl...
0626 2014 01_toronto-smac meetup_io_t
Bluegrass digital creds 2015
Introduction to NEW SAP - Accenture Technology Meetup
AWS Summit Berlin 2013 - Next-Generation Websites on Demand – with Infopark C...
Kafka: Journey from Just Another Software to Being a Critical Part of PayPal ...
DataStax on Azure: Deploying an industry-leading data platform for cloud apps...
About Tekmonks
Agoda open stack in a large scale deployment
SAP LEONARDO SAP LEONARDO the digital digital innovation innovation innovatio...
Power point updated
Application Services - CCS Technologies (P) Ltd.
meet4psa
Can Your Mobile Infrastructure Survive 1 Million Concurrent Users?
Digital Reinvention by NRB
Ad

More from WiMLDSMontreal (11)

PPTX
The Five Ws of Funding, by Sahar Ansary, Partner, R&D Partners
PPTX
The Agile methodology - Delivering new ways of working, by Sandra Frechette, ...
PDF
Coveo Machine Learning for E-Commerce: At the Center of Business Challenges, ...
PPTX
How to build a data science project in a corporate setting, by Soraya Christi...
PPTX
Diversity and Knowledge Production, by Jihane Lamouri, Diversity, Equity and ...
PPTX
Diversity & Deep Tech Start-ups, by Eleonora Vella, Program Director & Princi...
PPTX
Ubiquitous Machine Learning: Lessons from DeepRL in Robotics and Speech, by F...
PDF
Fashion-Gen: The Generative Fashion Dataset and Challenge by Negar Rostamzade...
PPTX
Artistic Applications of AI, by Luba Elliott, AI Curator
PPTX
What Scares Me About AI, by Rachel Thomas, Co-founder of fast.ai & Professor ...
PPTX
Using Feature Grouping as a Stochastic Regularizer for High Dimensional Noisy...
The Five Ws of Funding, by Sahar Ansary, Partner, R&D Partners
The Agile methodology - Delivering new ways of working, by Sandra Frechette, ...
Coveo Machine Learning for E-Commerce: At the Center of Business Challenges, ...
How to build a data science project in a corporate setting, by Soraya Christi...
Diversity and Knowledge Production, by Jihane Lamouri, Diversity, Equity and ...
Diversity & Deep Tech Start-ups, by Eleonora Vella, Program Director & Princi...
Ubiquitous Machine Learning: Lessons from DeepRL in Robotics and Speech, by F...
Fashion-Gen: The Generative Fashion Dataset and Challenge by Negar Rostamzade...
Artistic Applications of AI, by Luba Elliott, AI Curator
What Scares Me About AI, by Rachel Thomas, Co-founder of fast.ai & Professor ...
Using Feature Grouping as a Stochastic Regularizer for High Dimensional Noisy...

Recently uploaded (20)

PPT
Teaching material agriculture food technology
DOCX
The AUB Centre for AI in Media Proposal.docx
PDF
Approach and Philosophy of On baking technology
PDF
Profit Center Accounting in SAP S/4HANA, S4F28 Col11
PDF
Unlocking AI with Model Context Protocol (MCP)
PDF
7 ChatGPT Prompts to Help You Define Your Ideal Customer Profile.pdf
PPTX
Spectroscopy.pptx food analysis technology
PDF
MIND Revenue Release Quarter 2 2025 Press Release
PDF
Advanced methodologies resolving dimensionality complications for autism neur...
PDF
Review of recent advances in non-invasive hemoglobin estimation
PPTX
KOM of Painting work and Equipment Insulation REV00 update 25-dec.pptx
PDF
Electronic commerce courselecture one. Pdf
PDF
Chapter 3 Spatial Domain Image Processing.pdf
PPTX
Digital-Transformation-Roadmap-for-Companies.pptx
PDF
Architecting across the Boundaries of two Complex Domains - Healthcare & Tech...
PPTX
Programs and apps: productivity, graphics, security and other tools
PPTX
Detection-First SIEM: Rule Types, Dashboards, and Threat-Informed Strategy
PDF
cuic standard and advanced reporting.pdf
PDF
KodekX | Application Modernization Development
PPTX
Understanding_Digital_Forensics_Presentation.pptx
Teaching material agriculture food technology
The AUB Centre for AI in Media Proposal.docx
Approach and Philosophy of On baking technology
Profit Center Accounting in SAP S/4HANA, S4F28 Col11
Unlocking AI with Model Context Protocol (MCP)
7 ChatGPT Prompts to Help You Define Your Ideal Customer Profile.pdf
Spectroscopy.pptx food analysis technology
MIND Revenue Release Quarter 2 2025 Press Release
Advanced methodologies resolving dimensionality complications for autism neur...
Review of recent advances in non-invasive hemoglobin estimation
KOM of Painting work and Equipment Insulation REV00 update 25-dec.pptx
Electronic commerce courselecture one. Pdf
Chapter 3 Spatial Domain Image Processing.pdf
Digital-Transformation-Roadmap-for-Companies.pptx
Architecting across the Boundaries of two Complex Domains - Healthcare & Tech...
Programs and apps: productivity, graphics, security and other tools
Detection-First SIEM: Rule Types, Dashboards, and Threat-Informed Strategy
cuic standard and advanced reporting.pdf
KodekX | Application Modernization Development
Understanding_Digital_Forensics_Presentation.pptx

Building Analytics and Data Science at A Start-Up, by Kathleen Siminyu, Head of Data Science at Africa's Talking

  • 1. www.africastalking.com 1The Business of APIs Building Analytics and Data Science at a Startup
  • 2. Kathleen Siminyu Head of Data Science ● Mathematics and Computer Science background ● Self-taught data scientist ● Co-organiser of Nairobi WiMLDS (We turn 2 on Dec. 8th!) ● Organising Committee Deep Learning Indaba ● Head of Data Science at Africa’s Talking
  • 3. www.africastalking.com 3The Business of APIs Africa’s Talking is a Pan-African, Platform as a Service (PaaS) company based in Nairobi, Kenya. Our world-class infrastructure is created for developers and relied upon by businesses. We avail the complexity of SMS, VOICE, USSD, AIRTIME & PAYMENTS via simple API calls and at a fraction of the cost. ABOUT AFRICA’S TALKING One Continent. Unified platform. Unimaginable possibilities.
  • 4. www.africastalking.com 4The Business of APIs By The Numbers Active Countries Planned Countries 2018 Nigeria Ethiopia Côte d’Ivoire Kenya Botswana Tanzania Uganda Rwanda Malawi Zimbabwe Zambia DR Congo Sierra Leone Senegal Gambia Ghana South Africa Mauritius 11 Countries 20K+ Developers 30M+ API Calls Daily 5K+ Businesses
  • 5. www.africastalking.com 5The Business of APIs Airtime Pinless airtime top-up from 1 wallet across 4 markets SMS 2-Way Bulk and Premium SMS APIs across 6 Markets Payments One wallet for all your mobile, bank and card payments across 5 Markets USSD Unified Push/Pull USSD API across 6 Markets Build intelligent cloud-based Voice applications VOICE OUR APIs
  • 6. www.africastalking.com 6The Business of APIs Echo Mobile High impact mobile surveys via SMS, Voice, Airtime and USSD iShamba Using Voice & SMS to connect agriculturalists to experts in the field M-Kopa Customer engagement through SMS to provide affordable solar to off-grid homes CIC Bima Providing life insurance through airtime Naivas Bulk SMS for promotional content & customer engagement mShule Quality and personalized learning through SMS, USSD and voice USE CASES Over 5,000 business are currently consuming our services. Here’s how:
  • 7. www.africastalking.com 7The Business of APIs Session Overview • What We Did • Lessons Learnt
  • 8. www.africastalking.com 8The Business of APIs What We Did Technical assessment of our data situation - taking stock • Our core infrastructure is written in Scala • Analytics service powered by Redis, Cassandra and Akka Actors - mainly to service client dashboards • MySQL databases with client info • Different softwares used by finance, client relations, • Applications logs with everything else that didn’t make it into a DB
  • 9. www.africastalking.com 9The Business of APIs What We Did Explore opportunities and use cases for analytics and data science • Clustering of clients and markets • Time series predictions to predict traffic on our infrastructure • Anomaly detection, especially for our payment products • Building developer profiles using features like programming languages, errors, sophistication
  • 10. www.africastalking.com 10The Business of APIs What We Did Explore architectures and tools that might suit our needs • Settled on a lambda architecture ---> allows us to run both batch and streaming analytics jobs • Our stack • Apache Kafka • Apache Spark • HDFS from Apache Hadoop collection of softwares • Apache HBase • Scala - programming language of choice • Docker • Apache Mesos • Kubernetes • Apache Avro • Apache parquet
  • 11. www.africastalking.com 11The Business of APIs What We Did Build the team in-house • Hiring proved a challenge, almost no one in the ecosystem had experience with these tools • Changed tactics and convinced several engineers to slowly start porting their skills to become data engineers
  • 12. www.africastalking.com 12The Business of APIs Lessons Learnt • Have a dedicated team solely focused on data • Distributed systems can be a hairy situation esp security and logging • It’s important to manage expectations from the rest of the organisation • Buy in from the whole organisation is important to avoid duplication of efforts • Hold off hiring until you know what you are building and what kind of talent you need