HI, I have data on hourly mobile internet usage for each day and month (last 4-5 months). So this is basically a time series data for different customers. From this how can I understand whether few customers are offloading to wi-fi which is affecting the total mobile internet usage ? Any method or techniques ?
Your problem seems to be interesting. But before solving the specific problem I would love to see some trends of the known cases where consumer has switched to wifi and study their mobile internet usage changes. The approach after that could be -
- Simpler one is to observe the same trend in the new set of customer and mark one which behaves similar to the known customers as potential wifi switchers
- If you can create some profiles by doing cluster analysis, and see where the known cases were and where they moved after data offloading and observe the same movement with the new customer set and mark them as potential wifi switchers
Hope this helps.