As the 2012 election season approaches its inevitable climax on November 6th, the race between the two presidential candidates and their parties is heating up. The purpose of Chitika Insights’ Presidential Election Tracker is to provide data driven insight and highlight the level of online interest for Barack Obama and Mitt Romney. Here, the Chitika Insights team explains the methodology that powers the tracker.

Why did Chitika create the 2012 Presidential Election Tracker?

As the 2012 election season approaches its inevitable climax on November 6th, the race between the two presidential candidates and their parties is heating up. The purpose of Chitika Insights’ Presidential Election Tracker is to provide data driven insight and highlight the level of online interest for Barack Obama and Mitt Romney. We hope that it will serve as a way to keep your finger on the pulse of the political conversation and follow the progress of the respective candidates with the general electorate.

How do we calculate the data used in the Chitika 2012 Presidential Election Tracker?

Our 2012 Presidential Election tracker is powered with ad impression data coming through the Chitika Ad Network. To determine the respective levels of interest for each of the two candidates, their relative levels of strength in different geographic regions across the U.S., and their association with key issues, Chitika Insights is performing an analysis on the millions of search engine ad impressions from the Chitika Ad Network. The figures in each section are calculated across a minimum date range to guarantee statistical significance (see our expanded methodology for more detail).

Impressions:

Sample: This study was conducted across a full range of hundreds of millions of search impressions from the Chitika Ad Network; coming out of the U.S and Canada.

Determination: The intent of a user’s search query can be determined through analysis of the query string. All search impressions generally pass on query information in the form of a query string. Some examples of this are visible below:

Query: Obama and the economy

In this case, it would be reasonable to assume that this user demonstrated interest in Barack Obama, and the economy.

Query: Romney and the economy

In this case, it would be reasonable to assume that this user demonstrated interest in Mitt Romney and the economy.

Query: Obama and Romney debate

In this case, it would be reasonable to assume that this user demonstrated interest in both Barack Obama and Mitt Romney.

Reporting:

Top Line Graph & Accompanying Pie Chart: The percentages reported in the top line graph highlight the historical level of interest in Barack Obama and Mitt Romney on an hourly basis. The data is reported with a six hour delay to ensure maximum continuity. The accompanying pie chart relays the same information; except, instead of using a time series, the data is represented as an aggregate of the last 24 hours.

State Map: The state map highlights the level of interest driven by the two respective presidential candidates across the US, using a 24 hour average. The data is reported with a six hour delay to ensure maximum continuity. Percentages for each candidate are calculated based off of their level of search interest relative to the total impressions seen in that state. These percentages are then relayed across the map; states in which Barack Obama has a greater than 50% share of interest will be colored with a more blue hue, while states in which Mitt Romney exhibits a majority share of interest will be colored a more red hue.

Top Cities: The top cities section focuses on the level of interest driven by the two respective presidential candidates across the five largest cities in the US. This data is calculated using a 24 hour average with a six hour delay to ensure maximum continuity within the data. The percentage figure above the displayed city skyline shows the current level of search interest for the respective candidate, with the associated keyword displayed below.

Leading Cities: The leading cities for the candidates are calculated by measuring the search interest for the respective candidates across the 50 largest cities in the U.S.  This data is calculated using a 24 hour average with a six hour delay to ensure maximum continuity within the data. The two cities in which each respective candidate generate the highest ratio of search interest are then relayed on the 2012 Presidential Election Tracker.

Leading Issues: The leading issues for the 2012 election campaign are calculated using a list of keywords for each respective subject. This data is calculated using a seven day average with a six hour delay to ensure maximum continuity within the data. They are then dynamically ranked from one through ten, according to the level of volume generated by each issue. The size of the bar line next to each issue is driven by the corresponding level of volume generated.