Featured Data Project

Click Here to view the Dashboard

Data Visualization: Titanic Passenger Survival

by Prasad Pagade.


This project shows 3 different bar graphs. It shows percentage of survival for each gender, passengers travelling with family and location from where the passengers embarked the ship by passenger class. This dashboard helps reader understand what factors helped a passenger survive the fateful crash.


Exploratory Data Analysis and Cleaning (Python)

I downloaded the data from Kaggle – Machine Learning from Disaster, which was already clean and didn’t need extensive data wrangling or transformation. I used techniques learnt in P2: Investigate a dataset with Python and came up with few initial hypothesis. While exploring the data, I cleaned up some data like missing ages and embarked locations. I mapped the columns Pclass (1:Upper Class, 2: Middle Class, 3: Lower Class) and Embarked (S:Southampton, C:Cherbourg, Q:Queenstown) with more readable names so that it can be used directly in the charts later.

I created a new column person from the column Sex and Age so that we had three types of classification – Male, Female or Child(child is a person who is 15 years old or less). I saved this modified file as “titanic_data_modified.csv”

Data Visualization (dimple.js)

I decided to use dimple.js to produce the stacked bar chart and used d3.js objects to manipulate the html elements.

I considered using multiple chart types to test which one represents the findings most convincingly. I re-evaluated different chart type by tweaking few line of code and confirm my initial assumption, a bar chart is sufficient to display data.

Initial Design decisions

My initial designs included a stacked bar graph by gender. I wanted to include more graphs and animation from this point forward.

This initial iteration can be viewed at index_old.html, below are the screenshots few iterations of the graph:

First Chart

Second Chart

Second Chart

Third Chart


I gathered feedback from 3 different people and made iterations to my design:

Feedback #1

Great Visualisation.The graph conveys the message. it be easy for people to understand better, if we have a survival rate/percentage instead of people survived

Feedback #2

I think it is really good. It is “clean” and easily understood. While it is fine as it is, as we interact, the new category legends are added, but the old legends are not removed. I don’t know how much work that is to correct.

Feedback #3

The button position is wierd. On minimizing it overlaps on the chart. I would suggest fixing the button elements so that they donot overlap on the chart.

Post-feedback Design

I implemented the following changes:

  • Reduced the widths of the bar charts. So they look a lot cleaner.
  • Increased the font size of the legends and the axes.
  • Changed the Y-axis to percentage scale.
  • Fixed the button overlapping problem.
  • Fixed the legend clearing problem so that only the legends for the selected charts appear.
  • Added interactive text display which imparts user with information on each graph.

Final version of the data visualization is shown below:

Final Chart

Link to the Visualization



  • titanic_data_modified.csv: original downloaded dataset with minor cleaning.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s