Team: Noora Khorsandzadeh and Xiang Cheng
The Car Drive Project is about utilizing car drive data to contribute to social good in different areas, such disaster relief, events security planning, driving habits, and even urban environmental issues. We are attracted to this project because of the following reasons:
- Valuable and Big data: the data is car trip data of billions miles from Verizon; getting access to this kind of big and real-world data is exciting for data scientists.
- Real-world information: it is millions of drivers’ trip data; they are factual information reflecting how people mitigates.
- Numerous possibilities: we could combine the trip data with other datasets, such as weather data, local events, ect, and contribute insights in different areas.
Data Introduction:
The primary data set is collected from Verizon In-Drive devices on millions of cars. Each observation/row is information about one trip; the features includes:
- Trip Time: Start and End
- Vehicle Info: Make, Model, Year, Trim
- Miles Driven: Highway, City, and Total
- Fuel Consumed: Highway, City
- Speed: Avg, Max
- RPM: Avg, Max
- Battery Voltage with car on: Min, Max, Avg
- Battery Voltage with car off: Min, Max, Avg
- Check Engine Light on?: True or False
- Speed Bins (Counts): 0, 1-5, 6-10……>100 mph
- RPM Bins (Counts): 0-500, 500-1000….. >7000
The trip data mainly covers metro-Atlanta area (as shown in the figure below).
We are still requesting data sets from Verizon and expect to get them in 2 weeks.
Possible contributions to social goods:
In this project we are going to use the big data of billions of miles of car trips to find the models that can help improve the social good. One of our ideas is to use the weather data along with the car trips data to investigate the changes in the traffic of a specified region during a disaster, comparing to the normal day’s traffic. We believe that the result of this investigation can be useful for predicting the traffic model in a specific area while the weather forecasts show the similar conditions. So, emergency services can be organized better during the disaster.
Another idea is to use the car trip data to figure out the driving habit of the usual people, and the percentage of using highways instead of city roads. We believe that the result of this study would be important in deciding the road improvement plans in a specific region. So,the traffic flow would be improved, and driving for people would be much easier.
We may get more ideas after analyzing the data.
Future plan:
1. Research on published work:
There are already many published work using car trip data to study human behavior [1], traffic estimation [2], disaster relief [3, 4], etc. We hope to gain more ideas from these published work.
2. Other datasets searching:
As we mentioned in the previous part, We could explore a variety of areas if we have other information, such as weather, road, disaster, etc.
3. Preliminary data cleaning and analysis:
After getting the datasets, we will get familiar with them by preliminary cleaning, analysis, and some simple visualization.
4. Focused questions :
There will be many different questions we would like to answer from these data sets, but we have to pick a few focused questions for this summer.
5. Data processing, modeling, and visualization:
We will conduct in-depth analysis and modelling based the questions decided in the previous part.
6. Preparation of summary report and presentation:
We will summarize our results and share ideas for possible future work.
[1] Liao, Lin, Dieter Fox, and Henry Kautz. “Extracting places and activities from gps traces using hierarchical conditional random fields.” The International Journal of Robotics Research 26.1 (2007): 119-134.
[2] De Fabritiis, Corrado, Roberto Ragona, and Gaetano Valenti. “Traffic estimation and prediction based on real time floating car data.” Intelligent Transportation Systems, 2008. ITSC 2008. 11th International IEEE Conference on. IEEE, 2008.
[3] Zhao, Wenrui, Mostafa Ammar, and Ellen Zegura. “A message ferrying approach for data delivery in sparse mobile ad hoc networks.” Proceedings of the 5th ACM international symposium on Mobile ad hoc networking and computing. ACM, 2004.
[4] Montoya, Lorena. “Geo-data acquisition through mobile GIS and digital video: an urban disaster management perspective.” Environmental Modelling & Software 18.10 (2003): 869-876.