In this project, we aim to infer the quality of Project Sidewalk users based on their interactions with our system--both low-level interactions like mouse clicks and moves--as well as higher-level, more application related interactions (amount of panning on a street view image, etc.). More details are in our Dropbox folder ProjectSidewalk_PredictingUserQuality
Additional files you will need for the classifier (sidewalk-user-quality-analysis-neil.ipynb):
users_one_mission.csv
: CSV of quality information for users with at least one missionsidewalk-seattle-label_point.csv
: Information about labels in Seattleml-label-correctness-one-mission.csv
: Label correctness for all the labels of users with at least one missionall_users.csv
: Features generated from the feature generation notebook
If you want to be able to run the classifier from start to finish, use sidewalk-classification-notebook.ipynb
.