Our dashboard was very similar to our python app, but we took the feedbacks we received from our peers and TA's to further improve our R app. Some additions include, excluding the year 2010 that had no data in plot 3, better layout design, and adding in descriptions for our users to better understand the purpose of our app.
Overall, the app does well in answering our research questions of understanding unemployment rates across industries over a range of years. It also allows the user to interact with the plots in various ways with a radio button, year slider and multi-option select tabs using the many features that dash has to offer. Our dashboard is simple, yet effective, especially when we divided the graphs into 3 tabs. This allows the user to focus on one plot at a time and to gain insight with the single graph. Even with the limited amount of data we were able to display two different types of statistics in unemployment rates.
One of the improvements we noticed from switching from Python to R was that for our third plot we previously were not able to change the month numeric values to month names. With R, we were able to making it easier for user to understand what the values represent. We also decided to remove some tooltip options because we found that it is too redundant and is just repeating the information that the graph is giving us. Further, we added in an average line of unemployment across all industries for users to compare against the total with the respective industry that they are interested in.
Due to time constraints and the extent of our knowledge, there were a couple of known limitations:
- The changes to the slider are quite static, we hope for the future to improve the smooth transition when we use the slider to select different ranges.
- We also would have liked to add more features on our app that allow deeper interactions of the plots that allow users to understand the trends more clearly. For example, a detailed tooltip that could tell us more information about the country’s industry, but we were limited with the lack of knowledge on what country this dataset was from.
We used the Github issues to create a to-do list that let us communicate with each other the tasks at hand and made sure we were following our timeline and team contract.