histogram of traffic
Person's route through out the day
Heat map of position
We would like to have local business use this data to better market to their customers.
What it does
- Does a hotspot analysis of all the geolocation points in the month of September
- Since we feel points do not give you a good story because of the large amount of missing data, we decide to fill them in by giving an approximation of what they did in between. Using the data, we were able to trace out a route of where the person was at during any point during the day. We routed user from point A to B, not just a straight line from A to B
- Using the result from 2, we were able to construct a more accurate map of what the traffic is like at a certain point of the day
- Then we took a step further by analyzing the app usage data. We analyzed what is the most popular app used by people who travelled past a certain point. This will give business an idea of what kind of people travel by their business everyday. Then the business can use this data to target which social platform to focus their advertisement.
How we built it
Challenges we ran into
The routing problem was very challenging. The script did not work for various reasons, from incomplete data to mistake we made on our own.
We also had issues combining our result from analysis 3 and analysis 4 (see description above of what they are) As the data set was very large, we can only process so much data in such a short time frame, therefore, the dataset that intersected both 3 and 4 was very small.
Accomplishments that we're proud of
We are most proud of our analysis number 2. It was very difficult in getting the routing to work. Much the less routing through thousands of points.
What we learned
Data Science is hard.... We will not always get the results that we expect when performing these kind of analysis due to biases that we may have.