1. Build infrastructure for user uploads/downloads
   1. CSV
      1. AWS S3
   2. Output from Google Form
      1. MongoDB
   3. Geojson
      1. AWS S3
   4. Json
      1. MongoDB
   5. CartoDB dynamic updates?
   6. Jason
   7. make the data from the original sources
   8. store all datasets on S3 bucket
   9. Make a MongoDB that holds geojsons. Takes user input (eg PCode) and inserts a sub-dict into the object's json dict.
      1. Callout
      2. Callout
      3. Myanmar "Table"
      4. Townships "Table"
      5. Villages "Table"
   10. Accomplishments
       1. Completed Actions 1
       2. Completed Actions 2
       3. Completed Actions 3
   11. I don't know
   12. Action 1
       1. Action 1a
       2. Action 1b
   13. Action 2
   14. Action 3
   15. Action 4
   16. Action 5
   17. Action 6
   18. drag files & drop here
2. Discover segments/clusters of townships in the census data
   1. affluence factors
      1. cash
      2. banks
      3. credit cards
      4. cell phones
      5. electric power
      6. livelihoods/economics
   2. Anthony
   3. Who are the customers?
   4. We need to condense all of the census data features into a more concise set of features (PCA).
   5. Visualize and define clusters in a 2 or 3 dimensional state (kmeans, GMM, DBSCAN)
   6. Accomplishments
      1. Completed Actions 1
      2. Completed Actions 2
      3. Completed Actions 3
   7. Action 1
      1. Action 1a
      2. Action 1b
   8. Action 2
   9. Action 3
   10. Aug 20
   11. Action 4
   12. Action 5
   13. Action 6
   14. project sample document.pdf
3. Predict Ability to Pay
   1. Have a clear metric that enable decision makers to compare townships/villages (visually)
   2. Anthony
   3. PCA(census data matrix) => 1dimensional "affluence index"
   4. Random Forest/Linear Regression => Feature Importance
   5. Accomplishments
      1. Completed Actions 1
      2. Completed Actions 2
      3. Completed Actions 3
   6. Next Month
   7. scale all the features
      1. Action 1a
      2. Action 1b
   8. do PCA
   9. Interpret PCA vectors and feature scores
   10. Action 4
   11. Action 5
   12. Action 6
   13. drag files & drop here
4. Front End / UX
   1. Collect
   2. Describe
   3. Discover
   4. Predict
   5. Advise
   6. Cameron