Data exploration and processing of activity tracker data; users clustering and users analysis.
Here you find a short description of the project
Data Exploration:
-
Users: Visualize statistical profile of user data, filter noisy data.
-
Weights: Visualize statistical profile of weight data, filter noisy data, filter by metric (kilos), handle multiple measure at the same timestamp, compute bmi.
-
Activities: Visualize statistical profile of activity data, filter noisy steps counts based on threshold, steps vs distance ratio, and steps vs calories ratio.
-
Bmi, Steps: Merge bmi and steps data: each entry consists of user_id, measurement day (year, month, day), bmi, and steps. All steps entries are kept in the merged set. Missing bmi values are interpolated.
-
Bmi, Steps correlation: Correlations between bmi and number of steps considering all users, and inidividual users. Cross correlations bewteen number of steps and bmi for individual users.
Clustring and Retrieval:
-
Scripts: Python scripts to compute users clusters based on bmi, age and activity level; and to perform cluster based analysis for a given user (find best cluster, show individual results compared to similar users, show example of bmi vs steps evolution inside the cluster).
-
Samples: Sample reports for two different user data.