While it is quite nice aesthetically and cartographically speaking, the UI is quite hard to understand and use. For instance, the "click on or click-drag-drop" message should not need to exist: these are top level actions that the user should intuitively understand based on the overall design of the visualization tool. Moreover, "click-drag-drop" is confusing, as we aren't actually dropping anything (like of file into a folder), just doing a standard click-and-drag. Conversely, lower-level actions seem to be represented by icons and nothing else. For example, it took me a long time to figure out what the button that appears when you hover over the y-axis of the box-plot did (it resorts the x-axis based on their y-axis values, right?) . A simple tooltip would have solved that problem in less than one second.

This is not to disparage you specifically, but rather offer some constructive criticism: pay close attention to user interaction. It's an often overlooked portion of information visualization design, and most of us are guilty of ignoring it at one point or another. I know I sometimes get so caught up in adding all these neat features, bells, and whistles that I forget that somebody other than me is going to need to use it.

It's beautiful and effective; the biggest impediment to perceiving TEMPORAL trends is that time is the static variable in the display (until the user triggers a change, in which case they wait for an update). My guess is that you could pre-compute all of the variables for each year that you display, such that rather than re-rendering the graphic you would just flip an image, and caching would speed things up after the first viewing. That doesn't seem to be happening, which is why I presume that you're generating these on-the-fly or preventing caching in some other way.

Animating the whole graphic, or if that is too slow then allowing the user to animate the parts most interesting to them might support more temporal inference.

The comments are meant to be supportive - I think what you've done is fantastic - I just think the temporal aspect could be clearer.

Really nice interactive map. Very effective and easy to understand and navigate

These are fantastic graphics! I found that the interactive features worked smoothly for me and I was able to navigate it very well. It was like I new what kind of data I was going to find in the graphic even before I started playing around with it!

I do agree with the previous review that although prevalence data is easy to get from this, I think its a little harder to get a sense of trend. Perhaps a similar set of graphs focusing on trends? I realize this defeats the purpose of having it all in one place but it may help the user differentiate current state vs. changes over time.

Overall, great job!

Within a given year the visualization is very effective. I agree on temporal trends; how do I compare years? I'd love to be able to see 2 years side by side. Then the refresh delay would by less important.

While it is quite nice aesthetically and cartographically speaking, the UI is quite hard to understand and use. For instance, the "click on or click-drag-drop" message should not need to exist: these are top level actions that the user should intuitively understand based on the overall design of the visualization tool. Moreover, "click-drag-drop" is confusing, as we aren't actually dropping anything (like of file into a folder), just doing a standard click-and-drag. Conversely, lower-level actions seem to be represented by icons and nothing else. For example, it took me a long time to figure out what the button that appears when you hover over the y-axis of the box-plot did (it resorts the x-axis based on their y-axis values, right?) . A simple tooltip would have solved that problem in less than one second.

This is not to disparage you specifically, but rather offer some constructive criticism: pay close attention to user interaction. It's an often overlooked portion of information visualization design, and most of us are guilty of ignoring it at one point or another. I know I sometimes get so caught up in adding all these neat features, bells, and whistles that I forget that somebody other than me is going to need to use it.

Hi, It seems that this visualisation serves two purposes. On one hand it tries to provide a 'neutral' view on the data. But it also seems to want to draw attention to one particular 'feature': the sub-saharan cluster. I believe that you need to make a choice. This choice should be guided by the context in which the visualisation is shown and the purpose it serves. A visualisation can be seen as a speech-act in the context of a discourse. This speech-act has to make sense within this context.

Thanks Stuart for taking the time to review the visualization and give your feedback and recommendations.

As you realized, the user are able to use the slider (at the top-right corner of the map) to change the time (in years) and all visuals are rendered. The viz server is generating every view on-the-fly queering the live-connected data set and taking the advantage of cache to speed up the process for further user interactions.

I'm evaluating the option of adding a trend chart to clear the temporal aspect of the the visualization.

Thanks for your recommendation.

Best,

Ramon

Wow! While writing my comment to Stuard's answers, many other constructive and positive comments were added to this discussion.

Thanks to all for your feedback. I'm reading them carefully to improve this and further visualizations.

The RG community is fantastic!!

Best regards,

Ramon

Congratulations for your work.

I agree the interface is not intuitive.

Mostly, I would like to select a specific country and visualize temporal evolution of AIDS prevalence: is it possible?

Several countries data is unaivalable, in Europe and America.

Dear Ramon,

This is a very interesting and nice tool, congratulations on your work! As Stuart said, the highest impact of the visualization would be the temporal trends in each region/country, and that´s important to add. I suggest that you have a "temporal trend" tool that plots the overall HIV prevalence of a country over time every time the country is clicked on. You can use the space in the page occupied by the lower graphs, overdrawing them, instead of just graying them while they are not used.

I know this may be further down the road, but you can think of adding some molecular epidemiology data (HIV subtypes) in your site, although the only data we have is those by Joris Hemelaar at the WHO, which is discrete.

We also need to solve the problem of the HIV prevalence in most high income countries and in Brazil, since we expect an even higher rate of increase in prevalence in those countries. But that´s something beyound your scope...

All the best,

Marcelo

Dear Ramon: I agree that this is a great community. The quality of the input here reflects the quality of your posting - a high-quality question that reflects significant work and shows great potential is worthy of thoughtful responses.

I'm also happy to see such an excellent representation from around the world, including familiar names/faces! This is the Internet realized!

The interactive feature is very powerful. Are you writing the code for Tableau engine? Will it be available for use on a Mac? It would be interesting to be able to look at persistence with available anti-retroviral therapy over time .

Congratulations! It seems a very good work, but I am not an expert of this field (SIDA and AIDS). Having a temporal analysis of a single country may increase significance of this tool.

Hi Ramon, great job, very informative.

If I may add my 2 cents, on tiny details:

- the "click on or click-drag-drop.." sentence deserves to be highlighted or in another color/policy. At first quick look, it may be confused with a legend.

- legend: why did you use diamond shapes when outliers appear as rounds in the graphs?

It's really great that your graphs are dynamic and that clicking on outliers makes information popping up.

Congratulations

Hi Ramon, this looks great. I especially liked how you used the legend to give yet another comparison, making it easy to see how far apart countries are. Although understanding this would be slightly tricky and the small size of the legend might make it a bit hard to manage.

Hi Ramon, the tool is very nice and, at least fro people like me that need only key data, useful.

Thanks for all the suggestions and recommendations. I'm now implementing most of the suggestions.

William, Tableau is a commercial product by Tableau Software, you can find more info in its web site http://www.tableausoaftware.com

Current products are for Windows but a version for Mac is on development.

Dear Ramon:

Congratulations for your work. Let me suggest a couple of issues that I see could improve the visualization: first, lighten up the graph, eg allowing the reader to choose between displaying "distribution by regions" or "distribution by income groups". Moreover, for scale reasons, sometimes the box plot does not seem to be the best option in both types of graphics ...

In any case, a magnificent DataViz. Thanks for sharing it

Ramon,

Visually, your info graphic is appealing. As a whole, I enjoyed playing and learning from your graphic. The information, however, is quite dense and the user interface is not intuitive. This may be the result of my unfamiliarity with Tableau software. For an advanced audience I don't think it will present difficulties. If, however, your target audience is a general layperson, then the interactivity should be intuitively easy and rely on actions common to the layperson's general computer/software experience.

As mentioned by others, the temporal aspect of HIV was a piece I was was very interested in. The somewhat slow pause-load-show which occurs when changing the year was a let down. I was hoping to find an animate option to play the years in a loop form so I could study the changes in HIV prevalence and geolocations over time. Preferably, the year slider would have a much faster reaction time and an animate button. This allows the user to animate the time related changes, then manually step through the more relevant or interesting years. Also, the density of the material makes full screen visualization a necessity. My novice skills with the graphic were unable to find a maximize or fullscreen option.

I would like to end with a thank you. Thanks for a great data visualization and for making me aware of the Tableau software.

I have implemented some of the recommendations provided as answer to my questions.

This is an update of what I have done:

1.- Many has suggested to include the temporal aspect of the data, so I added it in a small line chart overlaying the left side of the map that is displayed once one or more countries are selected in any of the views of the visualization.

The chart is small because of space available on the visualization, but I think users are able now to get a better idea of HIV trends over time.

2.- I harmonized the type of symbol in map legend, replacing diamond by filled circle as in the box plots.

3.- I change some terms, such as click-drag-drop by the standard term click-and-drag

4.- I don't have much to do to speeding up the transition when user change the year and I would like also to implement a motion component, a playback type of control to show the data in motion but unfortunately the software I'm using does not provide that type of feature when the visualization is online.

For those interested, the visualization is available here: http://publichealthintelligence.org/content/what-percentage-people-ages-15-49-are-infected-hiv

Thank you all for your recommendations, I've learned a lot from all of you.

I expect to apply what I've learned in future projects

Best regards,

Ramon

Good work Ramón, it's very interesting and I have been impacted by the fact that no reliable data are available for such important countries as Canada, United States and Brazil among others.

Best regards

Emilia

Hola, Ramon:

Even if you job is nice because fo the tools you offer us, missing the information of so many important countries and regions is difficult to understand:

I will share with you a link of United Nations initiative for AIDs that give you the data about ALL the countries of athe world:

http://www.unaids.org/en/regionscountries/countries/ (HOME)

http://www.unaids.org/en/regionscountries/ (REGIONS AND COUNTRIES)

http://www.unaids.org/en/regionscountries/countries/unitedstatesofamerica/ (USA)

http://www.unaids.org/en/regionscountries/countries/canada/ (CANADA)

http://www.unaids.org/en/regionscountries/countries/ (WESTERN EUROPE)

Other: Andorra, Austria, Belgium, Denmark, Finland, France, Germany, Greece, Iceland, Ireland, Italy, Luxembourg, Netherlands, Norway, Portugal, Spain, Sweden, Switzerland, United Kingdom of Great Britain and Northern Ireland

http://www.unaids.org/en/dataanalysis/datatools/ (DATA TOOLS)

http://www.unaids.org/en/resources/infographics/ (AND HERE IS THE MAP OF PREVALENCE FOR ALL THE COUNTRIES OF THE WORLD)

For any future paper, research, etc I would be happy to cooperate with you,

Fermin de la FuentePh.D. B.Sc.

Senior Consultant

De La Fuente Consulting

Virginia Beach, VA 23451

e-mail: [email protected]