search

UBC-MDS / DSCI532-peer-review

0 stars 5 forks source link

olympic-dash - Group 22 #19

Open helingogo opened 2 years ago

helingogo commented 2 years ago
zackt113 commented 2 years ago

Reviewed by: Group 21

Overall, it is a very informative dashboard for the International Olympic Committee members. We like the way you presented data in four different graphics and provided a variety of ways to filter data. Our compliments and suggestions are as follow.

The peer review was done on Wednesday, March 9th. So, the suggestions were based on the work you have done before the time.

To complete by the group reviewing:

Persona

  1. In the scenario, it was mentioned that Harry would like to investigate if there are any abnormal performance trends for certain countries. We think it could be better if the dropbox of countries contains the full name of the country. Because some country codes could be hard to infer.
  2. As Harry would like to see the trend of different countries. We think the country variable should connect to all graphics. For now, the country variable can only filter the line chart. Connecting four graphs would make the dashboard much more interactive and informative.
  3. Having different filters for different graphs could be confusing at the first point. For example, the slider between the line chart and the density plot is ambiguous. Because there is no title of the slider and we have no idea whether the slider is connected to the density plot or the line chart. It would be great if you guys can add more annotations to filters and convert some local filters to global filters.
  1. As the dashboard was designed for International Olympic Committee members, they probably don't have sufficient time to wait for the dashboard to react. When I played around with different filters, the dashboard reacted fairly slow. We suggested that you may focus on some typical or more recent data to visualize in order to save computational power and running time. For example, having data in 1904 may not add a lot of value for today's decision-making.
  2. If the committee members would like to use the dashboard to deliver presentations, then it would be better if you guys can get rid of empty plots. For some selections and year ranges, there is no information for the scatter plot and bar chart. It looks a bit awkward when there are only two charts that have information to present.
  3. We think the interactivities on plots are insufficient. It would be better to add more things, such as interval selection, tooltips and click selections to make those plots more vivid. For example, if the committee member would like to show the number of gold medals earned by athletes, who are 27 years old, the graph cannot show the information directly, the audience has to guess from the height of the bar.

Frontend / layout

From our perspectives, the majority of those texts are easy to read, but some texts are too small and a bit ambiguous. Here are some suggestions:

  1. The title of the dashboard should be more obvious than normal texts. Adding background colour, changing font size or font colour may be helpful.
  2. We understand it is the tradition of plotting histograms, but for non-statistical professionals, the name of the y axis could be confused. Instead of using "Count", some more meaningful phrases could be used.
  3. The size of plot titles could be bigger as the current size is a bit smaller to us.

Yes, we think the selected plots are appropriate for the data and audiences. As we mentioned above, only the scatterplot has been implemented with some interactivity. More tooltips, interval selections, click selections could be added. Also, even better, if those graphs can interact with each other. For example, if the end-user select a bar in the bar chart, the density plot and scatter plot will highlight those athletes, whose height is between the selected range.

In general, it is easy to read in terms of layout. Just one little suggestion, it may be good if the title of the dashboard goes to the middle and above all graphs and filters.

We think it would be better if there is a tab or navbars. Because the information you guys would like to deliver is from very different dimensions, such as medal earned, athlete's height, athlete's age, year. Therefore, maybe graphics related to attributes of athletes could go to one tab and line charts could go to another tab. Even better, if you guys can come up with another chart, which illustrates more about the trend of medal earned, to stay together with the line chart.

Yes, most of the research questions were answered by those plots. Only one of the research questions was missed. (Which athletes have won the most medals across the years?). This question could not be answered by the dashboard. It would be great if the question could be removed or the dashboard can provide a table to list the name of the athletes, who have won the most medals across the years.

Backend

Those filters in the widgets are working fine. But the dashboard just react very slow, the suggestion was provided above. Moreover, there may be too many filters in the dashboard, which makes the dashboard a little bit overwhelming at the first point. Maybe focusing on a few filters would be better.

We don't think there are any two plots that are interacted with each other. And this would be an excellent area to improve on. Some specific suggestions were mentioned above as well. Also, the colour you used for the density plot looks perfect and is closely related to the topics. We think if the other plots can inherit these colors, the dashboard will become more interesting.

Documentation

Overall, documentation is complete and comprehensive. Here are some suggestions.

  1. Seems the sketch does not look similar to the actual dashboard. We assume it was probably because the team changed the plan afterwards. But for end-users, it may be confused. So, it would be better if the sketch could be updated.
  2. It looks like the key matrix part provides a brief summary of each graph. It would be great if the type of the plots could be specified in front of the short description. So, the users could grasp the main idea of each graph right away.

Just some minor suggestions, other than that, looks good!

We would say no because most of our group members spent some time getting familiar with the dashboard. And, the dashboard react fairly slow, so it was even more difficult to see the effect of each filter. And each graph also has its own local filter, which makes the dashboard become even more complicated. We totally appreciate the effort that you guys made on the dashboard, but maybe simplifying the dash a bit could make the dash be more concentrated on meaningful messages and be more user-friendly!

Indicate something you would like to include in your own app after reviewing this one

  1. We like how the group set the different sizes of the points in the scatter plot. We may consider implementing it into our own scatterplot as well to deliver more meaningful information.
  2. Our dashboard only supports two filters. Just like the group, we could add one more global filter to further divide information into another dimension.
  3. The unique colour of the density plot is very interesting. It not only self-explains the topic of the dashboard but also presents the medal-earned information more vividly. We would think about some great colours, which could potentially do the same job as the group!

Final final comment: Good job Team 22, it was our pleasure to go over your dashboard. And, please keep working on the dash as we are looking forward to seeing some more interesting graphics about Olympics!

Thank you, Group 21

ruben1dlg commented 2 years ago

Hello team! We addressed your feedback in the following ways:

  • Provide 3 pieces of constructive feedback that could help the app to be better suited to the scenario it was designed for.
  1. In the scenario, it was mentioned that Harry would like to investigate if there are any abnormal performance trends for certain countries. We think it could be better if the dropbox of countries contains the full name of the country. Because some country codes could be hard to infer.
  2. As Harry would like to see the trend of different countries. We think the country variable should connect to all graphics. For now, the country variable can only filter the line chart. Connecting four graphs would make the dashboard much more interactive and informative.
  3. Having different filters for different graphs could be confusing at the first point. For example, the slider between the line chart and the density plot is ambiguous. Because there is no title of the slider and we have no idea whether the slider is connected to the density plot or the line chart. It would be great if you guys can add more annotations to filters and convert some local filters to global filters.
  1. In this case, we believe that this should not be an issue, since the users are experts on the olympics, therefore they should be able to infer what country each country code refers to.
  2. Since we have another graph that is actually based on countries, it would not make sense to add this as a general filter. The bubble graph will only have one bubble if we add that filter.
  3. We addressed this issue in the following commit: https://github.com/UBC-MDS/olympic-dash/commit/9496b61044dae656e7b1e100cb3ca778ac23e3f1
  • Provide 3 pieces of constructive feedback that could help the app to be better suited to the user(persona) it was designed for
  1. As the dashboard was designed for International Olympic Committee members, they probably don't have sufficient time to wait for the dashboard to react. When I played around with different filters, the dashboard reacted fairly slow. We suggested that you may focus on some typical or more recent data to visualize in order to save computational power and running time. For example, having data in 1904 may not add a lot of value for today's decision-making.
  2. If the committee members would like to use the dashboard to deliver presentations, then it would be better if you guys can get rid of empty plots. For some selections and year ranges, there is no information for the scatter plot and bar chart. It looks a bit awkward when there are only two charts that have information to present.
  3. We think the interactivities on plots are insufficient. It would be better to add more things, such as interval selection, tooltips and click selections to make those plots more vivid. For example, if the committee member would like to show the number of gold medals earned by athletes, who are 27 years old, the graph cannot show the information directly, the audience has to guess from the height of the bar.
  1. We addressed this by preprocessing the data beforehand and reading the already processed data when opening the dashboard. This was addressed in the following commit: https://github.com/UBC-MDS/olympic-dash/commit/392c1ec628d4d949b740b30fad305a6a8529449d
  2. For this, we changed the year slider to only select valid years that would show information in the plots. This was addressed in the following commit: https://github.com/UBC-MDS/olympic-dash/commit/1b7065e66fbdbba9926845ab19dfc1529d5785b0
  3. We added tooltip annotations for some of the plots. This is the link to the commit: https://github.com/UBC-MDS/olympic-dash/commit/9496b61044dae656e7b1e100cb3ca778ac23e3f1

Frontend / layout

  • Is the text (plot labels / titles / subtitles) easy to read?

From our perspectives, the majority of those texts are easy to read, but some texts are too small and a bit ambiguous. Here are some suggestions:

  1. The title of the dashboard should be more obvious than normal texts. Adding background colour, changing font size or font colour may be helpful.
  2. We understand it is the tradition of plotting histograms, but for non-statistical professionals, the name of the y axis could be confused. Instead of using "Count", some more meaningful phrases could be used.
  3. The size of plot titles could be bigger as the current size is a bit smaller to us.

We addressed all three of this concerns in this commit: https://github.com/UBC-MDS/olympic-dash/commit/1b7065e66fbdbba9926845ab19dfc1529d5785b0

  • Are the plots selected appropiated for the data? Can you think of other alternatives or improvements for the plots regarding interactivity/reactivity or plot selection?

Yes, we think the selected plots are appropriate for the data and audiences. As we mentioned above, only the scatterplot has been implemented with some interactivity. More tooltips, interval selections, click selections could be added. Also, even better, if those graphs can interact with each other. For example, if the end-user select a bar in the bar chart, the density plot and scatter plot will highlight those athletes, whose height is between the selected range.

As referenced above, we added tooltips to some of the plots.

  • Was the dashboard organized in a way that was easy to read (columns / rows) or you would suggest a change in this sense?

In general, it is easy to read in terms of layout. Just one little suggestion, it may be good if the title of the dashboard goes to the middle and above all graphs and filters.

As mentioned above, we addressed the whole aesthetic of the dashboard 😀

  • Would you prefer including tabs / navbars to improve the navigation or you think it is better to visualize everything in one screen by the nature of the content?

We think it would be better if there is a tab or navbars. Because the information you guys would like to deliver is from very different dimensions, such as medal earned, athlete's height, athlete's age, year. Therefore, maybe graphics related to attributes of athletes could go to one tab and line charts could go to another tab. Even better, if you guys can come up with another chart, which illustrates more about the trend of medal earned, to stay together with the line chart.

We decided to stay with just one tab, since we only have 4 plots. We believe we would have very empty tabs if we used more than one.

  • Are the plots or other images described properly for the persona described?

Yes, most of the research questions were answered by those plots. Only one of the research questions was missed. (Which athletes have won the most medals across the years?). This question could not be answered by the dashboard. It would be great if the question could be removed or the dashboard can provide a table to list the name of the athletes, who have won the most medals across the years.

Backend

  • Are the options that you can select in the widgets working fine or have you found some cases where it is not working?

Those filters in the widgets are working fine. But the dashboard just react very slow, the suggestion was provided above. Moreover, there may be too many filters in the dashboard, which makes the dashboard a little bit overwhelming at the first point. Maybe focusing on a few filters would be better.

We addressed it by preprocessing the data, as mentioned above 😀. It got faster!, but nevertheless, we can see that in Heroku it is still a little slow.

  • If there are interactive plots: are the labels clear and easy to understand?

We don't think there are any two plots that are interacted with each other. And this would be an excellent area to improve on. Some specific suggestions were mentioned above as well. Also, the colour you used for the density plot looks perfect and is closely related to the topics. We think if the other plots can inherit these colors, the dashboard will become more interesting.

Documentation

  • Was the documentation easy to understand from the README.md how to use the app? How could the documentation be improved?

Overall, documentation is complete and comprehensive. Here are some suggestions.

  1. Seems the sketch does not look similar to the actual dashboard. We assume it was probably because the team changed the plan afterwards. But for end-users, it may be confused. So, it would be better if the sketch could be updated.
  2. It looks like the key matrix part provides a brief summary of each graph. It would be great if the type of the plots could be specified in front of the short description. So, the users could grasp the main idea of each graph right away.

Just some minor suggestions, other than that, looks good!

  • Was the app intuitive to use without reading the documentation? What information could be good to include?

We would say no because most of our group members spent some time getting familiar with the dashboard. And, the dashboard react fairly slow, so it was even more difficult to see the effect of each filter. And each graph also has its own local filter, which makes the dashboard become even more complicated. We totally appreciate the effort that you guys made on the dashboard, but maybe simplifying the dash a bit could make the dash be more concentrated on meaningful messages and be more user-friendly!

We added some additional annotations, changed the layout and added an 'about' section to explain everything a little bit better. We can see it in this commit: https://github.com/UBC-MDS/olympic-dash/commit/df755e39b48d8f25cf7f9f7cc18f4c8683fe11cc

Indicate something you would like to include in your own app after reviewing this one

  1. We like how the group set the different sizes of the points in the scatter plot. We may consider implementing it into our own scatterplot as well to deliver more meaningful information.
  2. Our dashboard only supports two filters. Just like the group, we could add one more global filter to further divide information into another dimension.
  3. The unique colour of the density plot is very interesting. It not only self-explains the topic of the dashboard but also presents the medal-earned information more vividly. We would think about some great colours, which could potentially do the same job as the group!

Final final comment: Good job Team 22, it was our pleasure to go over your dashboard. And, please keep working on the dash as we are looking forward to seeing some more interesting graphics about Olympics!

Thank you, Group 21

Thanks for the review Group 22!