This morning we talked as a group about when we wanted to tell Matthew (our contact at King County Metro) we would be available to present to the operations workers at King County Metro. He wants us to present to the people who would actually be using the software we are developing so they can give us feedback and ask any of the hard questions before we have to present to the higher-ups. In our discussion, we also decided to reevaluate where we were in the project, and talk about what still needed to get done for us to be on time with our deliverables and having them ready before the end of the program, which is only three weeks away! We decided to explicitly assign people to either the analysis or algorithm group, so that we wouldn't run into the problem of everyone trying to do everything as much. I decided I wanted to shift gears a bit and work with the algorithm and unit testing it to make sure all the functions fit together properly. To prepare for that, I did a bit more work documenting the code I had been using to do ugly rides analysis in R, so that the people doing analysis (Kristen and Kivan) could use it going forward. 
     In the afternoon we had a D3 tutorial, which was pretty cool. It's funny because before this I had seen D3 and Tableau as pretty equivalent tools in that I thought both of them had the main purpose of just helping the user create cool new, interactive visualizations. However, there are some major differences in the way Tableau and D3 are used, and they are no where near as similar of tools as I previously thought them to be. Tableau is very user friendly and easy to use, and it is a great tool for data exploration as well as creating visualizations. D3, on the other hand, has a very steep learning curve and requires the user to have a pretty good background in coding, and particularly JavaScript, HTML, and CSS. Users go to D3 when they know what visualization they want to make and they want to have ultimate control in specifying what it will look like and what it will do. Also, D3 has a lot of interactive capabilities that are nice on a web page because they allow readers to play with the visualizations and learn more in that way. Overall, D3 seemed pretty cool to me, if not also quite a bit harder to use than I had originally thought.

     A user came in today to talk to us about what it was like for him with his experiences getting around the city and having a mobile disability. He spoke to both the sidewalks and paratransit groups simultaneously, but it was unfortunate for us because I think he had a lot more insight that was relative to the sidewalks group instead of us. It turns out that he had never even used the Access buses and didn't have much feedback to give us about the system from a user's perspective. He seemed to believe that the system was for citizens who were more handicapped or had different disabilities than himself. 
    I also tried to get some of the ugly rides data into Tableau today so that I could start making some visualizations to better inspect what is going on. However, I immediately ran into trouble formatting the dates and filtering and showing the runs in some meaningful way. 

     Today's title doesn't really pertain to work in any way, but I was coincidentally also having a hard time coming up with a title to categorize today, and I wanted to give a shout out to her. Thus, the title is as stated above.
     Anyway, today was basically just another work day. I spent a good amount of time today talking over the check capacity script with Kristen. She's doing a really great job with it, and sometimes she updates me with what is going on and any problems she's running into. I like providing a second opinion and helping out in that way. 
    During my own work time I was still tackling the ugly rides situation. Nothing too novel to report. We also had a presentation today on tips and tricks for using social media data. I was hoping it would be a little more hands on, with her showing us some real data and how she got it and what she had to do to clean it and analyze it, but I feel like it was more about the basic concepts behind doing social media analysis. 

     Tuesdays start with team updates, so as always we just did a quick update about what we've been up to this past week. Basically we reported on how we have been working to get all the pieces of our rerouting function together, and then we also started creating the main function that will call each of the individual pieces. Our challenge coming up this week is going to be integrating all the different pieces and making them all fit together, as well as starting to think critically about how it will fit with the interface we want to give to the dispatcher. 
     The rest of the day we all basically just worked on different aspects of the project. I kept working on concerns with the ugly rides, but I also contributed to many discussions we had throughout the day about different little issues in our code and how we should address different problems that arose along the way. Ugly rides became much more annoying today when I started digging deeper, and realized that my sneaking suspicions about something being wrong with the cost per boardings seems to be correct. Therefore, I spent much of today revisiting old code, trying to make sure our original cost per boarding analysis for ugly rides was correct. To complicate matters though, the script that was actually used is on Frank's branch so I couldn't find it at first since I was looking in the master branch on github. Then, once I started testing to better understand the script and find any bugs, I realized that the csv file he originally used has since been updated with a csv that has different columns and column names, which voids different parts of the R file. In short, it was all just really annoying, and I'm still not sure yet if something is going wrong in the calculations, and if so, what is going wrong and how to fix it. 

      A  good portion of my day today was spent in various different meetings. In the morning, we had Thaissa Way come to present about Urban at UW. She spoke to us about the difference between looking at the micro and macro levels for a social good problem, and how understanding the history of an issue can help illuminate new solutions to addressing the problem today. She spent a good amount of time explaining the history of Gas Works Park, which is here in Seattle, and that was interesting to me because I didn't know anything about its history. However, I had a hard time seeing how her presentation related to our work, and which parts of her talk I could really apply to what we are doing. At the very end she did ask us about our projects and we talked a little bit about what each of us were doing.
     In my work time today I was looking at the ugly rides again, and specifically I was targeting the rides that ended up being $0 for a taxi ride, which is obviously a bug of some kind. Working with these anomalies, I noticed that a good number of them were referring to rides that ended up being cancelled or no-shows, which means they should not be counted in our data set anyways. However, there are still rides after that which are basically just messy and seem to have bad data, so I need to figure out how to deal with them. Looking at the specific data though made me reconsider if we are looking at the ugly rides in the correct way, and what will be most helpful to look at moving forward. 
     I also met with Anat and the TCAT group this afternoon to just give updates with what we're working on and where we are in those. The exciting thing with that is that the high schoolers are coming next week to do the circuit project we created, and I want to go help with that if I can. 
     After my TCAT meeting, we had reading group discussion. I did read the paper, but it was very wordy and I guess I must have zoned out a bit, because I really did not understand completely what its purpose was entirely. The discussion leader asked us a lot of questions, but I felt that I didn't participate much and I think he thought that most people didn't read at all. I felt bad because I did read, but I guess the material just wasn't as interesting to me. I don't have much background in the Millennium Development Goals, or what health is like in Africa (Tanzania in particular), so that might be why I felt I couldn't or didn't want to contribute as much. I liked the discussion more than last week, but I guess I just thought the content wasn't very relevant to me or our projects.