The Real-Time Crime Forecasting Challenge seeks to harness the advances in data science to address the challenges of crime and justice. It encourages data scientists across all scientific disciplines to foster innovation in forecasting methods. The goal is to develop algorithms that advance place-based crime forecasting through the use of data from one police jurisdiction. This Challenge will offer a comprehensive comparative analysis between current "off-the-shelf" crime forecasting products used by many police departments and more innovative forecasting methods used by other scientific disciplines.
Winnters Announced: 06/30/2017
Uber is opening up in an area where it might make sense competitively for it to stay more closed off: The ride-hailing company’s new Movement website will offer up access to its data around traffic flow in scores where it operates, intended for use by city planners and researchers looking into ways to improve urban mobility. Uber says it was looking at all the data it gathered and began to realize that it could be used for public benefit, and assembled a product team to make this happen. Essentially, according to Uber, it’s hoping to make it easier for those with influence over a city’s transportation picture to make the right decision, and to be able to explain why, where and when the changes are happening with accurate data backing them up.
More info: Click here
We wish to extend our congratulations to Zohreh, a graduate student of the BDLab, for being chosen as the Outstanding Graduate Student of the computer science department. The award is well deserved. We can't wait to see what else Zohreh achieves.
Siddhant, one of the members of the BDLab, graduated with an M.S. for his work on a Multivariate Spatiotemporal benchmarking system. He has continued on to TransUnion where he is now a senior data analyst. We wish him the best of luck!
Our BD Lab team of Rob Fitzgerald, Nivin Alexis Lawrence, and Max Lees developed the prototype for an application to improve the experience of flow chart drawing, titled "Chartographer." The final application used AWS to host a Python-based Support Vector Machine, with a radial basis function kernel. It was trained on a data set which we manually produced via our mobile client interface, which written using OpenGL ES (Embedded Systems) library, in Java using Android Studio. Users can draw common shapes on a phone or tablet surface, and those shapes will be classified and re-drawn in a more perfect representation. This allows for the rapid expression of org charts, database diagrams, and molecular diagrams, for example. HackCU 3 took place on April 22nd and 23rd, on CU Boulder Campus. It was an overnight hackathon event, with awards for innovative use of various hardware and software solutions. It was attended by 387 students from Colorado institutions and beyond. Many caffeinated beverages were consumed.
Details of our project can be found at our submission web page.
Please join us to congratulate Shahab Helmi, our Ph.D. student, who had his paper, titled "Spatiotemporal Range Pattern Queries on Large-scale Co-movement Pattern Datasets", accepted for publication in the highly prestigious IEEE Big Data 2017 conference (acceptance rate: 18%). We wish him many years of great achievements.
More information about his paper can be found at shelmi.com.