Tuesday, June 30, 2015

Day 2

Today I looked at some of the D3 tutorials and I tried to figure out the code for some of the sample graphics. I was able to understand some of the simple ones and get them to work, but considering the complexity that will be required for visualizing the data, I needed to find another way. At least for now.

So I did some research for good software that I could use to create graphs and maps. After some experimentation with various resources, I found one that I particularly liked. With Tableau Public I created 3 visuals using some USGS data about the impact of Hurricane Sandy. This tool allows me to represent data in an appealing and easy-to-understand way with many different formats to choose from. It doesn't require any code, so I'll be focusing on becoming more familiar with all the tools and finding the best way to represent different types of data. I had some trouble modifying the labels on the axes and I haven't decided how I will handle large data sets just yet, so I'll just have to keep working at it.

Each of the bubbles represents a different region in New Jersey or New York hit by Hurricane Sandy. The relative size is the extent of the damage to businesses in each area. You can see the total number of businesses hit by hovering over a bubble with your mouse. 

Graph 2:
Here I focused on accommodation and food service businesses (such as hotels and restaurants) damaged by flood waters in each of the regions surveyed by USGS. I chose to do a sideways bar graph because it's fairly easy to see how the numbers varied from region to region. Hovering over each bar provides the specific values.

Graph 3:
This is a bar graph for Atlantic City. Each colored bar represents a different type of business, and its length is the amount of businesses affected. The x-axis labels are in alphabetical order, unfortunately, so it's not as easy to see how the different bars compare to one another. I'll have to go back and fix that.

3 comments:

  1. Nice work, Sasha! Now a few suggestions to make your blog posts more fun to read:

    1. Use hypertext, it rocks! For example, when you write "With Tableau Public...", You can make "Tableau Public" a link to: https://public.tableau.com . That would save me (and all your readers) the trouble of searching for it myself, and make reading your blog faster and easier. Similarly with all of the links to your visualizations. Make them active links so we can click on them.

    2. Use lots of visuals. In addition to linking to your visualizations, you could include a few screen shots of them. In fact, you could include a screen shot that is also a hyperlink to the real visualization.

    On a side note, I'm sitting here at my grandmother's house, 15 miles from Atlantic City, so your topic of Hurricane Sandy is near and dear ;-)

    One last thing, when it comes to mapping things on the earth's surface, OpenStreetMap (https://www.openstreetmap.org) is the tool of choice. I plan on starting my own visualization investigation by following the LearnOSM tutorial (http://learnosm.org/en/) either today or tomorrow, and I'll be blogging about that, so we can compare notes.

    ReplyDelete
    Replies
    1. Thanks for your feedback! I fixed the links and I'll make sure to include lots of visuals in future blog posts. I was thinking about working with USGS Hurricane Sandy data from a large Excel file that I found on their site (http://pubs.usgs.gov/ds/0905/). My idea is to create a sort of visual story with lots of different types of graphs and maps. I'll definitely check out OpenStreetMap. It looks like a great tool!

      Delete
  2. You have done some impressive work in such a short period of time! This is exactly what I was hoping you can show with these free tools.

    I definitely agree with jelkner's comment to try and create screenshots of these visuals you are referring to in this and future posts. It will look really neat to see the visual progress you are making throughout your internship just by scrolling down your blog site.

    Instead of OpenStreetMap, which is more about helping to geographically map places, I think you would enjoy exploring CartoDB (https://cartodb.com/gallery/) and Plotly (https://plot.ly/) to create interactive maps and graphs.

    ReplyDelete