DAVLectuer3 Exploratory data analysis .pdf
- 1. THE IMPORTANCE OF DATA VISUALIZATION IN BUSINESS INTELLIGENCE CS 352 – Data Analysis and Visualization 1
- 2. Topics covered Chapter 1 Introduction 2 Why Do Modern Businesses Need Data Visualization? The Future of Data Visualization How Data Visualization Is Used for Business Decision-Making Introducing Data Visualization Techniques
- 3. Shifting from Input to Output Chapter 1 Introduction 3 Visual Business Intelligence (BI) Tools →take quick decision In modern BI tool: Output → well structured and presented Driven by customers demands Vis Dashboards → highlight business performance and areas to be improved
- 4. Why Is Data Visualization Important? Chapter 1 Introduction 4 Dynamic interpretation of BI data using graphs Interactive and appealing dashboards → serve the business insights Easily figure out → patterns, trends, and correlations in the data Present a complete overview of the data Help in making better, quick, and informed decisions
- 5. Why Do Modern Businesses Need Data Visualization? Chapter 1 Introduction 5 TO ANALYZE HUGE VOLUME OF DATA GIVE DESERVED IMPORTANCE TO AREAS/COMPANIES/CUST OMERS GENERATING MORE REVENUE/IMPROVED PRODUCTIVITY THE BIRTH OF BUSINESS INTELLIGENCE DETECT INFO, RELATIONSHIPS, TRENDS AND PATTERNS WITHIN DATA. PREDICTING UPCOMING TRENDS/SALES/REVENUE MARKETS FIND CORRELATION BETWEEN OPERATING CONDITIONS AND BUSINESS PERFORMANCE UNDERSTAND CUSTOMERS BEHAVIOR AND INTERESTS
- 6. The Future of Data Visualization Chapter 1 Introduction 6 Data Vis:Was an Art ------> Now is Science Network Theory: Employs algorithms to understand and model pair-wise relationships crime prevention, disease management, social network analysis, biological network analysis etc. IoT: Analyze and visualize data streams collected from billions of interconnected devices (smart appliances, smart cities monitors and so on). Other Applications: complexity theories, nanoscience, social science research, education systems, cognative science, space, and much more
- 7. How Data Visualization Is Used for Business Decision-Making Chapter 1 Introduction 7 Faster Responses: Huge Data, understand, interpret →quick decision Simplicity: Complete Picture, Simplified View → consider essential data only Easier Pattern Visualization: Identify Trends, Patterns → efficient decision and advance planning Team Involvement: diff teams, same data, different perspective/goals → collaborate to achieve strategic goals Unify Interpretation: Different tools, same graphs → improved collective understanding
- 8. Introducing Data Visualization Techniques Chapter 1 Introduction 8 Aim: Extract graphing info to understand data, show patterns, spot trends, and identify outliers etc. Two ways of data visualization: Exploration: Extract info collected data Explanation: demonstrates → extracted info Many Types: temporal, multidimensional, hierarchical, and network
- 9. 1. Matplotlib 2. Seaborn 3. Plotnine(ggplot) 4. Bokeh 5. pygal 6. Plotly Python Data Visualization Libraries for Business Analytics 9 7. geoplotlib 8. Gleam 9. missingno 10. Leather 11. Altair 12. Folium
- 10. 1st and the oldest Python data visualization library Few libraries like pandas and Seaborn are build over matplotlib Learn more: www.matplotlib.org Allow to create all basic graphs Matplotlib 10
- 11. Seaborn uses matplotlib and takes fewer lines of code. Its charts → more aesthetically pleasing and modern Learn more: http://web.stanford.edu/~mwaskom/software/seaborn/index.h tml Seaborn 11
- 12. Plotnine is a python implementation of ggplot2, an R plotting system powerful visualization package using layer components to create a complete plot Tightly integrated with pandas, so it's best to store your data in a DataFrame when using Plotnine. Learn more: https://plotnine.readthedocs.io/en/stable/index.html Plotnine (ggplot2) 12
- 13. Plotnine (ggplot2) 13 Change in Rank (Plotnine
- 14. Create interactive, web-ready plots → integrate as JSON objects, HTML documents, or interactive web apps. Supports streaming and real-time data. Provides three interfaces with varying levels of control to accommodate different user types Learn more: http://bokeh.pydata.org/en/latest/ Bokeh 14
- 15. Create interactive, web-ready plots → integrate as JSON objects, HTML documents, or interactive web apps. Supports streaming and real-time data. Provides three interfaces with varying levels of control to accommodate different user types Learn more: http://bokeh.pydata.org/en/latest/ Bokeh 15 Interactive weather statistics for three cities (Bokeh)
- 16. pygal also offers interactive plots that can be embedded in the web browser ability to output charts as SVGs (Scalable Vectors Graphics) With huge dataset → trouble rendering and become sluggish. Learn more: http://www.pygal.org/en/latest/index.html pygal 16
- 17. pygal 17 Box plot (Florian Mounier)
- 18. Plotly as an online platform for data visualization Access its capabilities from a Python Offers some unique charts → contour plots, dendograms, and 3D charts Learn more: https://plot.ly/python/ Plotly 18
- 20. toolbox for creating maps and plotting geographical data Map type charts → choropleths, heatmaps, and dot density maps Dependency → Pyglet Learn more: https://github.com/andrea-cuttone/geoplotlib geoplotlib 20
- 22. allows to turn analyses into interactive web apps using only Python scripts, don't need to know any other languages like HTML, CSS, or JavaScript Works with → Any python vis lib Learn more: https://github.com/dgrtwo/gleam Gleam 22
- 23. Gleam 23 Scatter plot with trend line (David Robinson)
- 24. allows you to quickly gauge the completeness of a dataset with a visual summary filter and sort data based on completion or spot correlations with a heatmap or a dendrogram. Learn more: https://github.com/ResidentMario/missingno missingno 24
- 25. missingno 25 Nullity matrix (Aleksey Bilogur)
- 26. Basic level charting but quick and easy SVGs can be extended to larger datasets Learn more: https://leather.readthedocs.io/en/latest/index.html Leather 26
- 27. Same as seaborn but it depends on Vega and Vega-Lite Good for creating interactive visualizations easily and quickly Learn more: https://altair-viz.github.io/index.htm Streamgraph (Altair) 27
- 28. python library and mapping capabilities of leaflet.js (a Javascript library) visualize geospatial data can build a variety of interactive maps such as choropleth maps, scatter maps, bubble maps, heatmaps, etc it’s various plugins like Markercluster, ScrollZoomToggler, DualMap that let you wrap leaflet maps and extend its functionality. Learn more: https://github.com/python-visualization/folium Folium 28
- 30. https://mode.com/blog/python-data-visualization-libraries Chapter-2: Book - Data-Analysis-and-Visualization-Using- Python Further Reading 30
- 31. What is the aim of data visualization? Explain two ways of data visualization. Chapter 1 Introduction 31 The Aim of Data Visualization is to extract graphing information to understand data, show patterns, spot trends, and identify outliers etc. There are two ways of data visualization: Exploration: In this type of visualization we extract information from the collected data Explanation: In this type of Visualization we demonstrates the extracted information with various graphs