This was my first submission to Vis, and our team was remote. Yifang and I were in HKUST while Ke was in MSRA. It was a fresh experience in which I improved my research and front-end coding skills fast but also found my shortcomings to address.
Both Ke and Yifang are experienced and reliable. Working with them is lucky and comfortable. During this submission, I truly saw the power of teamwork. There is no way to complete all the jobs within three months by myself alone. Ke is a doer, decisive and highly effective. When the design requirements were not clear and data were deficient, he didn’t trap in problems but decided to build the prototype first and find some open source datasets. Yifang designed the style of the system, which I personally think is elegant and harmonious. She pays attention to every detail and pursues perfection. I appreciate their guidance and learned a lot from them, especially their personal qualities.
A lesson I learned is that in remote collaboration it is hard to keep everyone at the same pace as we cannot communicate immediately. I think possible solutions include a higher meeting frequency, daily report and keeping Skype opening the whole day when the deadline is close.
My front-end coding skill improves a lot in this submission from my point of view. Here are some important tips I learned:
Apparently, I need to make much more efforts in my writing. My biggest problem is that I attempted to explain everything in the paper but forgot the logic and some common rules of writhing visualization papers.
When I read related surveys and papers, I found that visualization really solves the pain points of anomaly detection. For instance, anomalies are rarer and have various forms, so it is hard to define and diagnose them only using algorithms. The evidence of the effectiveness of visualization in anomaly detection is strong. This inspires me about my choice of the PQE topic. If my research is about using visualization solve other questions, does visualization straight hit the pain point and is it irreplaceable? If my research is about developing a visualization technique such as data movie, could this technique serve for real application problems and has broad influence?