rawgraphs / rawgraphs-app Goto Github PK

Trial data:
name,x,y,size
A,0.1,0.1,30
B,0.9,0.9,20

Problems:
a) smaller bubble is (practically) not shown (too much in the corner)
b) bubbles with extreme x or y are cutted (not whole bubble is shown)
c) axes' labels are cutted, particularly the y-axis, but the x-axis sometimes, too

Hi, I've installed the local version and the python server works fine. The only issue is that the index page does not display the box/form to drop the data. There are links as well but which directs to no-where, for instance http://localhost:4000/#layout.
Is that an installation problem or there is something missing. I'm running Ubuntu Precise 12.04. Many thanks

Write tutorial

Following our talk of sooner today, having a layered .svg output would be a huge help for working on the result in Illustrator. Cheers!

Using pretty urls like /documentation instead of /?page_id=101

http://i.imgur.com/pQ9BCsg.png?1

bower -v 1.2.7
node -v v0.10.18

I did change the port to 8080 not sure if that could affect anything though

Hi there,
Do you mind sharing which app was used to generate the animated tutorials?
Thanks!

Hi,

please consider to add axis limits management: if I add a point to (3,5) another one to (10, 10) and finally another to (1000,1000) please let us define we want axis X and Y go from 0 to 15.

Thanks in advance
Bye
Piero

After updating to OSX Maverick, trying to open the SVG i'm getting the following error:
"The operation cannot complete because of an unknown error. [CANT]"

Something seems a little odd in sizes of the bubbles in bubble charts. In the example in these screen shots, the value of the largest bubble is 251.1 (Mississipi in the top left corner) while for the smallest it's 119.4 (Minnesota) or an easier one to see, Hawaii at 134.7, but the disparity in bubble size for these values is grossly over-exaggerated. The values should correlate to the area of the circles in my understanding.

The Data:
State Obesity Rate Physical Activity Heart Disease Death per 100000
Alabama 67.7% 42.4% 236
Alaska 64.8% 58% 151.5
Arizona 62% 52.4% 146.7
Arkansas 68.7% 45.8% 222.5
California 60.3% 58.2% 161.9
Colorado 55.7% 61.9% 132.8
Connecticut 62.3% 52.8% 155.7
Delaware 66% 48.7% 175.7
District of Columbia 51.9% 57.7% 222.4
Florida 62.1% 52.9% 162.3
Georgia 64.6% 50.8% 192.6
Hawaii 56% 58.6% 134.7
Idaho 62.5% 57.3% 159.3
Illinois 64% 51.7% 181.7
Indiana 65.5% 46% 191.8
Iowa 64.7% 47.7% 173.3
Kansas 65.6% 46.8% 164.9
Kentucky 66.9% 46.9% 210.1
Louisiana 69.6% 42.1% 229.4
Maine 64.2% 56.8% 151.1
Maryland 63.8% 48.8% 182.2
Massachusetts 58.8% 56.5% 150
Michigan 65.6% 53.6% 204.2
Minnesota 63% 54.1% 119.4
Mississippi 68.9% 40% 251.1
Missouri 65.7% 49.6% 201.8
Montana 61.3% 55.4% 154.2
Nebraska 65% 49% 154.2
Nevada 62.5% 52.8% 197.3
New Hampshire 62.1% 56.2% 152.7
New Jersey 61.6% 53.4% 182
New Mexico 62.8% 52.3% 151.2
New York 60.6% 51.6% 199.9
North Carolina 65.8% 46.9% 174.9
North Dakota 66.2% 47.5% 158
Ohio 65.3% 51.7% 192.4
Oklahoma 67.8% 44.8% 235.2
Oregon 61.2% 61.2% 137.9
Pennsylvania 64.9% 49.5% 187
Rhode Island 62.9% 48.9% 167.1
South Carolina 66.1% 50.1% 189.9
South Dakota 66.1% 46.2% 155.2
Tennessee 65.4% 39% 217.4
Texas 65.1% 48.3% 181.1
Utah 57.8% 55.9% 143.2
Vermont 60.3% 59.3% 153.6
Virginia 63.6% 52.5% 168.5
Washington 62.3% 54.3% 151.5
West Virginia 68.3% 43.1% 211.2
Wisconsin 66.4% 57.5% 165.1
Wyoming 63.3% 53.3% 169.8

string with date element at the end or at the beginning is parsed as Date

date sample:
label bug_label_1 bug_label_2
this_is_a_label_string this_is_also_a_label_string_2012-02-02 2012-02-02_this_is_also_a_label_string

I would add the "show points" checkbox in the options.
user can switch on/off the actual data points both on hexbin and on scatterplot.

Elements are not ordered by anything.

No alphabetical order, no order by value, not ordered as they are in the dataset... if I re-upload data and try again, they change order but still without any sense.

Opening the app, without pasting any kind of data, it already says "Ok, everything seems fine with your data. Plese continue".
I know it is correct (no data = no errors) but it should appear only when you paste data.

As we noticed during the test with the users, the "don't drop here" is not really visible during the operation of mapping. I suggest to make the text red or at least more visible.

Hi,

please consider to add some spacing around graphs.

Thanks
Bye
Piero

Safari 5 on Mac OSX Snow Leopard

Trying to download svg, it downloads the file without name and without extension.

Write FAQ for end-users

Write GNU/GPL license

Hi! I wanted this morning to deploy an instance of Raw on an address like http://mydomain.com/raw, but it did not work because every imports are written with absolute paths (at least here, for instance).

Could it be possible to switch them to relative paths? Thanks in advance!

Rercords with 0 (zero) as Link Value generate a mismatch between total amounts in different Dimensions.

As you can see in the picture there is a different height between the second step and the third one which is not supposed to be. the second step and the third one are supposed to be identical (padding is set to 0). If, in the records I said, I set the Link Value to 1 (one) instead of 0 (zero) it works perfectly.

Raw identifies the decimal part only if separated by point "."
When the decimal part is marked by comma, Raw discard it, keeping the integer number.
In my opinion or Raw parses it as float, keeping the decimal, or it stores values as strings.
This is because keeping just the integer part can bring to conceptual errors: the user is sure that values are correctly parsed while they are not.

What about giving the possibilty to map the size of the circles?

Can we map the size of the each element using a value instead of the count?
Can we put labels? I suggest to put them radially like in the circular dendogram aligned towards the center.

Put simply, this is a pretty awesome application. However, it seems like it's power is more in its abilities than as a stand-alone site -- at least that's where my interest lies.

I do realize that since the code is all here, one could pull out whatever's needed and form their own library to work with the code on a programmatic level, but I was curious if there was any intention of this being officially supported? E.g. install the Raw module, call some importData(...) method, apply mappings, set the rendering options, and then saveAs("svg") or similar?

Again, pretty impressive project, regardless!

Find the data sample

typo in the sentence:
"Ok, everything seems fine with your data. PLESE continue"

This thing is AWESOME, I've already got it connected to my Neo4j DB and am doing some great stuff with it. It'd be great if you could add a title within the actual visualization, to provide context to the data. Should be fairly straight forward to accomplish.

when rolled over, labels often do not display the /size/ (quantity), sometimes they will.

weird results if one or more values are 0.

When I install locally, requests to many of the non-application files are getting the index.html file as a response.

When going to localhost:4000 -- the index loads, and I see the title "RAW" and a bullet list of "data, layout" etc... None of the css or js is loading. When I inspect the request for /angular-ui/build/angular-ui.min.css with chrome inspector, the response looks just like the one I get from the request to /
(ie~ starting with: <!DOCTYPE html><html ng-app="raw"><head><meta... )

You can reproduce with: docker run carver/raw (assuming you have docker installed)

Which is basically a:

git clone <raw repo> /opt/raw
cd /opt/raw
npm install -g bower
npm install
npm start

Using commit 0ede2ee

On larger datasets and visualizations (such as the Titanic sample data + Alluvial Diagram) changes to the form inputs can cause the page to hang, even before the new value shows in the form field.

I think implementing something like Underscore.js's debounce() function [1] on these fields would make this much smoother, and even eliminate a lot of processing entirely when using the step controls or typing a multi-digit number.

[1] http://underscorejs.org/#debounce

in alluvial diagrams, the output graphic is weird if the decimal part is divided with a comma

Some elements have to be disabled under certain conditions (e.g. disabling export if no vis yet...)

After choosing the data, clicking any of the nav links at the top (Layout, Mapping, etc.) such as demonstrated in the video, jumps the focus to the top of the page and clears the data -- essentially bringing the app back to its initial state.

Tested with http://app.raw.densitydesign.org/ and locally, and with current versions of Chrome and Firefox. Haven't noticed any messages in the console log.

Raw recognizes dates, but when you use them as numbers (like in a scatterplot, or a bubble chart) it returns only the day number in the mont as value, regardless the mont and the year.

date
01/03/2001
02/03/2001
03/03/2001
04/03/2001
05/03/2001
01/04/2011
03/12/2013

Viz created with the sample dataset "Cocktails". Cocktails and Ingredients as Hierarchy. Parts as Link Value. padding set to 9.

Need a key or legend to export when creating a tree diagram.

Want: Input Type and Quantity data and convert into a Tree Diagram
Error: Doesn't automatically use all ten lines of data, uses only 9

Want: Key to reflect the colour and prioritise Quantity data in key (legend)
Error: Can only export Tree diagram as a .png, not the key too

Want: the colours to be different enough so can work with 100 lines or more of data
Error: Not enough differentiation for a non-builder to distinguish between Type categories

Otherwise, easy to use and quick!
: )

I would keep "scatterplot" instead of "bubble chart" (bubble chart is ambiguous)

Many data have geo-references. Add a way to visualize spatial data with lat, lon using an OSM map.

Awesome tool!! Can we have a heat map visualization?