Hi!
Over the past few months, we were able to redevelop the original audit log analyzer incorporating the feedback we received on this forum. This time, we implemented our project using Vue.js to align with the ODK Central codebase, and added some new features that we wanted to get feedback on. This redeveloped version has a better user experience and more readable graph displays.
Link to Repo:
https://github.com/UDub-Impact/ODK_Vue_Audit-Log-Analyzer
Here are the new features:
- More detailed user interface that includes specific instructions on how to get the audit log from sandbox and upload the data to the analyzer.
- Added error checking for uploaded forms so that users are able to see why the CSV is not able to be parsed. Our previous analyzer did not provide any feedback making it difficult to tell why exactly an uploaded CSV was being rejected.
- Top bar displays general statistics about the form such as number of questions, number of responses and average response time for each submission
- Users can also now sort, filter data and display the top K or bottom K results, making it easy to recognize outliers.
Future development plans:
Our goal with allowing the users to focus on viewing the top/bottom K features was to handle forms with a large number of questions. We assumed that individuals using the analyzer would primarily be focused on outlier values, and thus we decided that providing this flexible limit would ensure that the graph had a reasonable amount of bars. However, we were thinking that it could still be valuable to provide the option to display all the data for all the questions on large forms if that is something users would benefit from. Would this be a beneficial feature despite the possibility of the graph being difficult to read due to its size? We would love some input on this!
Questions for the community:
- We are considering adding a feature where a user can select the specific questions that are included in the analyzed charts displayed. Does this seem useful, and is there anything else that you would like to see in terms of limiting the amount of data used in the analysis?
- Are there any suggestions in terms of the flow and design of the analyzer for ease of navigation?