Visualize data before choosing a machine learning algorithm
- Identify patterns
- Find corrupt data
- Identify outliers
- Find imbalances in the data
- Explore and demonstrate important relationships, and strength of relationships (density), using plots/chats
Charting data
Types of information to convey via Business Intelligence (BI) tools
- Key Performance Indicators (KPIs)
- Relationships
- Comparisons
- Distributions
- Compositions
Key Performance Indicators (KPIs)
A single value that represents a particular area or function and shows relative performance
- Net Promoter Score (NPS)
- Customer Profitability Score (CPS)
- Conversion Rate
- Relative Market Share
- Net Profit Margin
Use KPI charts to represent these indicators
Relationships
Establish or prove a relationship between 2 or more variables
Best chart to use depends on number of variables being compared
- Scatter Chart
- Two variables, example: social media spend to adoption rate
- Bubble Chart
- Three variables, example: comparing investment return, investment duration, and investment commitment
Comparisons
Show how variables change over time or show a static view of how different variables compare
Best chart to use depends on number of variables being compared
- Bar Chart
- One variable, example: website hits in a given month
- Table
- Three variables, example: two dimensions represented as the columns and rows, the third by the data in the cells
- Column Chart
- One or two variables changing over time, example: show year-over-year sales and number of marketing campaigns
- Line Chart
- Three or more variables changing over time, example: show year-over-year sales, number of marketing campaigns, and web traffic
Distributions
Show how data is distributed over defined intervals, Interval meaning clustering or grouping, not time
- Column Histogram
- One variable, example: showing how many voters are in various generation groups
- Counting something and putting them into buckets
- Scatter Chart
- Two variables, example: relating return on investment, investment duration, and investment size
- X-axis is investment time, y-axis is return on investment, and the bubble size is the investment size
Compositions
Show the elements that make up data set, static or changing over time
- Pie Chart
- simple share of total
- Stacked 100% Bar Chart
- components of components
- Tree Map
- share of total
- Stacked Area Chart
- 5 or more periods
- relative and absolute differences
- Stacked 100% Area Chart
- 5 or more periods
- relative differences
- Stacked Column Chart
- less than 5 periods
- relative and absolute differences
- Stacked 100% Column Chart
- less than 5 periods
- relative differences
Business Intelligence Tools
- Amazon Quicksight
- Cloud powered business intelligence service
- Create interactive dashboards that include machine learning insights
- Anomaly detection, forecasting, auto-narratives
- TensorFlow with TensorBoard
- Tableau
本文由
Oscaner
创作, 采用
知识共享署名4.0
国际许可协议进行许可
本站文章除注明转载/出处外, 均为本站原创或翻译, 转载前请务必署名
