Redesign Averages Chart Interface

[jkoschinsky]

Several people have found it difficult to navigate the averages chart interface and found the output too complex/confusing (especially for the regimes+ time option). Here is a mockup for a proposed redesign of the interface that Luc and I discussed today (see link to PDF below mockup for more details). We can skype about the details if anything is unclear.

The idea is to use a dialog similar to that for the correlogram, which remains open after the view is generated so people can change the parameters on the fly.

avg chart dialog.pdf

[lixun910]

Please verify in 1.7.99

[jkoschinsky]

Feedback on the new interface has been really positive.

One question: For selected at time 0-1, "excluded" and "all" are also highlighted in the graph, which I think is confusing. Is it possible to only highlight selected?

Same with excluded:

and with selected vs. excluded, which also highlights all:

[jkoschinsky]

Can you add the variable name back in the header of the view? Right now it only says: Averages Comparison Chart (see correlogram and other views)

[lixun910]

Please verify in 1.7.101

[jkoschinsky]

It looks like the fix only displays selected or excluded trends if either are chosen for different time periods, which works, too. For this case (selected-excluded, same time), all is also highlighted in yellow. Is there a way to only yellow-highlight selected and excluded but not all?

[lixun910]

Sure, I leave it as a reference line on purpose. I will fix this in next build.

On Mar 28, 2016, at 6:01 PM, jkoschinsky notifications@github.com wrote:

It looks like the fix only displays selected or excluded trends if either are chosen for different time periods, which works, too. For this case (selected-excluded, same time), all is also highlighted in yellow. Is there a way to only yellow-highlight selected and excluded but not all?

https://cloud.githubusercontent.com/assets/13679743/14094825/ca717202-f50e-11e5-982c-32eef8f09d6b.png — You are receiving this because you were assigned. Reply to this email directly or view it on GitHub https://github.com/lixun910/geoda/issues/381#issuecomment-202650337

[lixun910]

Please verify in 1.7.103

remove average line; update “Apply” button to “Run Diff-in-Diff Test”; remove checkbox “Run Diff-in-Diff Test”; update some warning messages; update “Save Dummy” functions;

[jkoschinsky]

Luc: It seems that the bar for the overall mean is missing in the graph (see below). This used to be a horizontal line in black.

Julia: I think we lost this in response to one of the comments in ticket 381 re. only using the yellow highlight for e.g. selected, instead of highlighting selected, unselected and all for selected. Ideally, we'd put the overall mean back in and still only highlight what is selected instead of all three.

Luc: The overall mean should definitely be in, it gives perspective to the other values.

[jkoschinsky]

Looks like the overall mean is still missing?

[lanselin]

in a pure cross-sectional setup, it is there (just tested on Cleveland)

On Apr 7, 2016, at 8:11 AM, jkoschinsky notifications@github.com wrote:

Looks like the overall mean is still missing?

[Uploading Screen Shot 2016-04-07 at 5.10.57 AM.png…]()

— You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub https://github.com/lixun910/geoda/issues/381#issuecomment-206843296

[jkoschinsky]

The problem might be related to not having made a selection in one of the views but choosing selected vs. excluded in the averages chart. In that case GeoDa treats the overall mean as excluded:

Sometimes you also get a zero line for selected, which wasn't visible in the above view:

[jkoschinsky]

no, these were defaults after changing selection

On 7 April 2016 at 10:10, Xun Li notifications@github.com wrote:

See the vertical axis in the first figure: it’s from [1.8 - 7.4] — that’s why the selected is not shown. Did you change the Y axis range?

On Apr 7, 2016, at 5:18 AM, jkoschinsky notifications@github.com wrote:

The problem might be related to not having made a selection in one of the views but choosing selected vs. excluded in the averages chart. In that case GeoDa treats the overall mean as excluded:

< https://cloud.githubusercontent.com/assets/13679743/14350681/a60bd74a-fc7f-11e5-87e8-af90d34d2086.png

Sometimes you also get a zero line for selected, which wasn't visible in the above view:

< https://cloud.githubusercontent.com/assets/13679743/14350757/116a45a8-fc80-11e5-8868-195ba56d0a92.png

— You are receiving this because you were assigned. Reply to this email directly or view it on GitHub < https://github.com/lixun910/geoda/issues/381#issuecomment-206847225>

— You are receiving this because you authored the thread. Reply to this email directly or view it on GitHub https://github.com/lixun910/geoda/issues/381#issuecomment-206997723

---

Julia Koschinsky, Ph.D. Research Director Associate Research Professor Arizona State University School of Geographical Sciences and Urban Planning GeoDa Center for Geospatial Analysis and Computation julia.koschinsky@asu.edu

http://geodacenter.asu.edu http://www.facebook.com/geodacenter http://twitter.com/GeoDaCenter

[jkoschinsky]

Let's just put the black line for all back into selected-excluded (and, if possible, only use yellow highlighting for the chosen option, i.e. either selected or excluded instead of the current highlighting of selected, excluded and all). If nothing is selected just have zeros for the obs-mean-sd values and only show the line for all.

[jkoschinsky]

When I talk to people about the "Averages Chart" there's some initial confusion about its purpose.

To clarify this upfront, what do you think about renaming it to the "Difference-In-Means Chart" or to "Difference-in-Means Tests"?

[lanselin]

I actually like “averages chart,” nobody else has it :-) how about differences chart? i don’t want it to sound too complicated, four words is definitely too many.

On Apr 11, 2016, at 11:38 AM, jkoschinsky notifications@github.com wrote:

When I talk to people about the "Averages Chart" there's some initial confusion about its purpose.

To clarify this upfront, what do you think about renaming it to the "Difference-In-Means Chart" or to "Difference-in-Means Tests"?

— You are receiving this because you commented. Reply to this email directly or view it on GitHub https://github.com/lixun910/geoda/issues/381#issuecomment-208406598

[jkoschinsky]

Then let's keep Averages Chart since people won't otherwise know what the differences are about :)

On 11 April 2016 at 09:46, Luc Anselin notifications@github.com wrote:

I actually like “averages chart,” nobody else has it :-) how about differences chart? i don’t want it to sound too complicated, four words is definitely too many.

On Apr 11, 2016, at 11:38 AM, jkoschinsky notifications@github.com wrote:

When I talk to people about the "Averages Chart" there's some initial confusion about its purpose.

To clarify this upfront, what do you think about renaming it to the "Difference-In-Means Chart" or to "Difference-in-Means Tests"?

— You are receiving this because you commented. Reply to this email directly or view it on GitHub < https://github.com/lixun910/geoda/issues/381#issuecomment-208406598>

— You are receiving this because you authored the thread. Reply to this email directly or view it on GitHub https://github.com/lixun910/geoda/issues/381#issuecomment-208443621

---

Julia Koschinsky, Ph.D. Research Director Associate Research Professor Arizona State University School of Geographical Sciences and Urban Planning GeoDa Center for Geospatial Analysis and Computation julia.koschinsky@asu.edu

http://geodacenter.asu.edu http://www.facebook.com/geodacenter http://twitter.com/GeoDaCenter

[jkoschinsky]

Verified in GeoDa 1.7.107 (OSX Yosemite, 10.10.5) Data: Test Data.zip

Steps:

1. Choose time grouped variable
2. Make selection on map

Result:

[jkoschinsky]

Verified in GeoDa 1.7.109 (OSX Yosemite, 10.10.5) Data: Test Data.zip (same as before)

Steps:

1. Open map, then averages chart
2. Choose time grouped variable (BnkAll time 0-time16) in averages chart and leave selected vs unselected
3. Make selection on map

Problem persists:

[jkoschinsky]

GeoDa 1.7.109 (OSX Yosemite, 10.10.5)

Steps:

1. open time-grouped variable with select/unselect but no selection: red zero line is displayed (right)
2. if you make a selection and then unselect, the Y-axis is truncated above zero, so you don’t see the red zero line (left)

Since the red zero line might be confusing, why don’t we set the default for 1 to that of 2 or to some value above 0. I played with this using 0.1 and 0.2 but in this case the red zero line shifts below the x-axis instead of not being displayed:

[jkoschinsky]

GeoDa 1.7.109 (OSX Yosemite, 10.10.5)

It looks like in the process of dealing with the highlighting issue, we lost the trend line that connects the averages for each time period (or may be we never had it). Let's add this back in (same as black line for selected/unselected) and leave the highlighting in brown and purple as it is now.

[jkoschinsky]

The time/all problem is fixed and the Y-axis drop below X, too.

Now the only remaining problem is still related to the Adjust Value Range of Y-axis right-click option: For instance, in the example below, if I want to adjust the Y-axis to 2.1 should be able to do this but am no longer able to do this and asked to specify a Y value of zero or below:

The other thing is that after you edit the Y-axis, if you click on it again, instead of giving you the view to edit the Y values, it jumps back to the default setting.

[jkoschinsky]

User can now specify Y axis range that's narrower than the value range but that's easy to recognize and fix since you don't see anything. Closing for now.