Tableau Practice: Los Angeles Water Usage

Playing with a little Tableau using open data from Los Angeles Open Data on “Residential water use by month averaged for fiscal year. Numbers represent Hundred Cubic Feet (HCF) of water use.”

Couple of observations with the data:

  • Water usage seems to correlate with residential density, eg. the high water usage in Downtown and Central City that have more high-density apartment blocks versus suburban tract homes that make up most of Los Angeles.
  • Water usage dropped dramatically in Fiscal Year 12/13, I’m almost wondering if that data was never fully populated. Either way, there seems to be a good downward trend which is good because of the CA drought and just being a region where water is scarce.
  • Tableau’s sum filters can be puzzling for those who aren’t familar with the platform when you filter the data – eg suddenly the scale is off so you have to reduce that filter to see the values.

A few process notes

  • I had to pivot the raw data a bit before bringing it to Tableau
  • I imported a custom shape for the water drop effect
  • I used the area chart for trend overtime to kind of hit home the point the amount of water usage rather than a simple line to show time series.

Click here or on the visual below to take a look:



Why I Can’t Find Recycling Bins in NYC (visualized)

Although I love NYC, as a Californian transplant the one big pet peeve I have other than the lack of Mexican options and fresh avocados is the lack of recycling and general awareness of doing so. Yes, I’m that Californian who brings the reusable bags to the market and carries a water bottle at times in NYC.

I couldn’t help noticing walking around how many people buy items like bottled water and have no place to recycle them even though NYC has a lot more public waste baskets than a lot of other cities. Taipei, I’m looking at you and all those days I carried my trash for miles.

I played with some of the NYC Open Data in Tableau and did confirm how few Public Recycling Bins are in NYC.

For some reason they’re particularly absent in Queens, outer Brooklyn, and Upper East Side. Population density seems to have no impact. Shrug, come on NYC, you “progressive bastion,” get your recycling on! It kills me seeing how many people here buy beverages in containers that could easily be recycled and toss them into a regular trash bin. It’s crazy.

Click here or the image below to see the visualized data:


Major League Soccer 2015 Team Goal Stats



First go at this dataset.  I created an interactive dashboard-like view just looking at goal breakdowns by Major League Soccer teams overall for the 2015 season.  Planning to add several views to this panel with the data.


As per verbatim from my previous post:

I’ve really into soccer in general, especially international play, after my grad school project where I worked at the Annenberg Innovation Lab collaborating with Havas Sports and Entertainment and IBM on a research project studying soccer fans (see Fans.Passions.Brands, Fan Favorites, and Sports Fan Engagement Dashboard). I also did my degree practicum on Marketing to Female Sports Fans.

I’m now in another universe creating data visualization at an advertising agency and am trying to combine the geeky fandom with practical practice related to my daily work.

On that note, most of the statistics and visuals I’ve found through a just a cursory look are about wins and losses.  I’m trying to show goal data by team in the MLS in a way that looks at performance based on other factors such as number of attempts and assists and not just the win-loss-draw type metrics I found in most of the soccer sites I saw.


I deliberately tried to use Tableau components and styling that was out of the out-of-the box template for the platform to emphasize the ability to size values on an additional data dimension. For instance, sizing each bubble based on number of goals or customizing labels out of the default. I notice a lot of users of Tableau don’t deviate much from the standard template, and I’m trying to train myself to go beyond that and also get better at the aesthetic piece of data viz.


I use Tableau at work on a daily basis. I personally think where Tableau shines is its exploratory data capabilities if you know how to prepare data in a form usable in Tableau. A few years ago, its explanatory data visualization capabilities were second to none in this space, but the desktop tool has lost some flash factor to D3 and HTML5 visuals, but definitely not substance in my opinion. Plan to expand on this with win-loss figures as well as analysis of kicks as the high number of goals by team didn’t necessarily line up with number of matches won.

Data Viz Practice: NYC SAT Score Explorer

I made the following visualizations below to show:

  1. Score averages and number of test takers by District and Borough, with ability to search data tables for specific school information
  2. Look at distribution of scores on the three sections, with ability to filter by Borough, District, and School Name.
  3. Look at score correlations of the three sections with ability to drill down by Borough, District, and School Name as well as scroll test score.

Data Sources:

The Tableau embed below doesn’t seem to work well on WordPress viewing in Chrome: link to the Public Site here.

District Summary.png

Score Distributions.png

Critical Reading and Math Correlation.png

Day 63 of 180 Days of Data Viz Learning #jfdi

I’m doing some form of data visualization learning for 180 days because I need to #JFDI.  See post explaining how and why I’m doing this.

Decomposition of a Visualization:

Vox US GDP Comparison Time

  • What are the:
    • Variables (Data points, where they are, and how they’re represented):
      • A-axis time
      • Area graph slider for GDP change through time
      • Countries by globes with map symbol
      • Individual country GDP metadata in tooltip
    • Data Types (Quantitative, Qualitative, Categorical, Continuous, etc.):
      • Qualitative, Continuous
    • Encodings (Shape, Color, Position, etc.):
      • Position, Shape, Symbol
  • What works well here?
    • Showing trends through time – > eg seeing effect of dip in US GDP during depression
  • What does not work well and what would I improve?
    • Hard to tell what countries are what other than tooltip
    • Would be more compelling if certain events in history were annotated, eg depression, wars
  • What is the data source?  Do I see any problems with how it’s cited/used?
    • Gapminder
  • Any other comments about what I learned?
    • Really clever way of using animation and symbols

Code Learning:

Tableau Javascript API Tutorial

I’ve actually watched all the videos in this series on the Tableau Training site, which are probably some of the most well-done training videos I have.  This interactive tutorial is a nice brush-up for me and is better organizing than their docs for implementing work – in my opinion.

Three Takeways:

  • There are three categories of functions to select/interact with views, filter functions, switching tabs, and selecting values.  The distinction is important for development and not that clear upfront (at least it wasn’t to me).
  • Methods generally different between different even handlers that take in single versus multiple values
  • Filter
    • REPLACE method to only show one filter value
    • ADD allows multiple filter criteria
    • REMOVE clears values from filter

Reading and Learning Data Visualization Theoretically/Critically:

Reading Nathan Yau’s book Data Points – Visualization that Means Something

p. 68-87 Entertainment, Data Art, The Everyday

Three Takeaways

  • Nathan Yau quoting Amanda Cox “All of the great chartmakers make me feel something: alarm, wonder, surprise, joy … something.  Even, I think you might argue in the case of something like dashboard design, calm.” p. 69
    • I want the BI dashboard work I do to make people feel calm
  • Great example of drawing out simpe visual with humorous expression: p. 71
  • Data art is a way to allow yourself to “immerse yourself in the data, which is both personal and easy to relate to” p. 81

Day 39 of 180 Days of Data Viz Learning #JFDI

I’m doing some form of data visualization learning for 180 days because I need to #JFDI.  See post explaining how and why I’m doing this.

Decomposition of a Visualization:

Note to self:  Haven’t been doing these as much lately, but incredibly helpful for learning how to communicate and to get ideas.  Try to get at least three of these a week so you put in the extra thought to what you see and read.

Infographic: Fewer Americans Are Seeing the Benefits of Higher Education

  • What are the:
    • Variables (Data points, where they are, and how they’re represented):
      • Students enrolled in degree problem denoted by size and color
      • Students dropping out denoted by angle
      • Time on x-axis, debt and funding y-axis in area “mountain” view with different colors
      • Time on x-axis, number of students foreign and American on y-axis in “mountain” view with different colors
    • Data Types (Quantitative, Qualitative, Categorical, Continuous, etc.):
      • Quantitative
      • Categorical (students)
      • Continuous (time)
    • Encodings (Shape, Color, Position, etc.):
  • What works well here?
    • Use of size angle and flowing shapes very compelling in showing proportions of students and what happens
  • What does not work well and what would I improve?
    • Two bottom mountains graphs appear to be connected because of size and color but are not.  I would change the the color scheme more to denote the difference.
    • Color schemes in general are too similar for different types of data, making a compelling story a little less so because of extra processing time needed to understand the visual.
  • What is the data source?  Do I see any problems with how it’s cited/used?
    • Not clearly cited.  Not good.
  • Any other comments about what I learned?
    • Interesting use of combining size, angle, and position to position a psuedo-linear story.

Code/Technical Learning:

Hodgepodge of tasks:

Three Takeways

  • Javascript Review on Treehouse
  • Integrating D3 with Tableau
  • Getting better at SQL queries to create efficient tables to work with

Day 38 of 180 Days of Data Viz Learning #jfdi

I’m doing some form of data visualization learning for 180 days because I need to #JFDI.  See post explaining how and why I’m doing this.

Code Learning/Technical Learning:

Tableau Javascript API Training

Three Takeaways:

  • How to deal with date times
  • Filtering for min and max
  • How to set-up for exporting to PDF

Reading and Learning Data Visualization Theoretically/Critically:

WSJ Guide to Information Graphics

Chapter 3 Ready Reference

Three Takeaways

  • Log scales allow you to include values that span orders to magnitude on x-axis and adjust grid lines so incremental change at different values of the y-axis reflects its relative significance.  p. 100-101
  • Watch out for using non-comparable scales – readers expect a flat line for small increases and a steeper slope for bigger increases.
  • Charing the absolute values or the percentage changes from the initial data point yields the same shape for the graphs.  The chart plotting the percentage changes accentuates the changes from the baseline. p. 107