Appendix A: Data Visualization - Fraud Detection Using Machine Learning

Appendix A: Data Visualization

You can visualize the transactions this solution processes using Amazon QuickSight or your own visualization tools. Use the following procedure to configure Amazon QuickSight.

  1. Open a text editor and copy the following code.

    { "fileLocations": [ { "URIPrefixes": [ "https://s3-us-east-1.amazonaws.com/bucket-name/" ] } ], "globalUploadSettings": { "format": "CSV", "delimiter": ",", "textqualifier": "'", "containsHeader": "false" } }
  2. Modify the URIPrefixes URL.

    • Modify the region name to match the region where you deployed the solution, if necessary.

    • Modify the Amazon S3 bucket name to match the name you specified for Results Bucket Name AWS CloudFormation template parameter during deployment.

  3. Save the following code as manifest.json.

  4. Navigate to the Amazon QuickSight console. .

  5. Select Manage Data.

  6. Select New data set.

  7. Select S3.

  8. For Data source name, enter a name. For example, fraud_detection_events.

  9. For Upload a manifest file, select the Upload radio button and click the folder icon. Navigate to the manifest.json file you saved earlier.

  10. Select Connect.

    After a few minutes, a success message should appear that shows that the S3 data imported into the Amazon QuickSight in-memory calculation engine (SPICE). SPICE acts as a cache for the data stored in your data source. For more information, see SPICE in the Amazon QuickSight User Guide.

  11. When the Finish data set creation window appears, select Edit/Preview data.

  12. On the Edit page, change the column headers to more meaningful names. For example, Timestamp, ID, Source, and Fraud.

    
          Amazon QuickSight SPICE dataset

    Figure 2: Amazon QuickSight SPICE dataset

  13. Select Save & Visualize

Now, you can create graphical representations of your data. For more information, see Working with Amazon QuickSight Visuals in the Amazon QuickSight User Guide