Modern Metrix

Surveillance Cameras as a Source for Consumer Research Data

Think Google Analytics for brick and mortar stores, using in-store surveillance cameras:

"Prism's software can count people that come into a business, measure the length of the line at checkout, and produce static or animated visualizations showing how people moved around a store. It is designed so that it cannot identify or track individuals. One national wireless carrier is already using Prism's technology to generate heat maps of where visitors go in their showrooms, to compare the level of interest in different devices—valuable data to them and to the device makers."

I personally believe that data collected this way is more reliable than when using mobile phones to log and track people's behavior inside the stores.

How accurate are those cardio machines when it comes to calorie counting?

I've always regarded with sceptitism those "numbers of calories burned" reported by the cardio machines at the gym, but my guesstimate of the measurement error was usually 5-10%. Today I finally went online to see if I could find more information. According to this source, the error is much higher in some cases:

Treadmill: Overestimated calories burnt by 13 percent
Stationary Bike: Overestimated calories burnt by 7 percent
Stair Climber: Overestimated calories burnt by 12 percent
Elliptical: overestimated calories burnt by 42 percent

From now on, forget the elliptical machine!!!

Klout vs. Peer Index Scores: analyzing differences

While searching for some articles on influence, I stumbled upon this blog post by Christopher Penn, where he was challenging readers to dig into a small dataset of Klout and PeerIndex score to explore the differences between these two metrics of influence. I quickly ran some numbers which I included in my comment to his post, but would like also to share here.

The dataset that Chris was kind enough to post contains Klout and PeerIndex scores for 15,737 Twitter users who tweeted at least once in the past month, presumably on topic related to #Marketing (data is available on his blog, in CSV format). I first computed straightforward Pearson's correlation between two scores, which turned out to be .45. The rank-order correlation (correlation between the ranks of the scores instead of their actual values) was slightly lower (.43). Finally, I assigned the scores to quintiles (5 equal groups; think of them as top 20% of users with highest scors, then next top 20%, etc., ending with bottom 20%) and looked at the correlation and cross-tabs between the quintiles. The correlation was .42, the average overlap between two scores (that is, user who was in the top 20% on Klout score wa also in the top 20% on PI score was about 30%). And here is a modified confusion matrix for all 5 quintiles, showing the overlaps between the groups:

As you can see, the top and bottom 20% had higher overlaps. Overall, there is a positive relationship of moderate strength between two metrics. But only moderate.  One possible explanation for discrepancies is that Klout and PeerIndex use different reference groups. Remember, the scores are computed, normalized and compared within a larger group of users. Since it is nearly impossible to download and process data on ALL Twitter users, so most likely, these services draw random samples of people on Twitter and then use them as reference groups. 

Four new and growing "disruptive" business sectors

This article in 'Enterpreneur" discusses current "hottest" growing business sectors. Among the most "disruptive" categories, in my opinion, are the following four:

Decision Support: recommendation and discovery engines and platforms, particularly, when it comes to consumption and sharing of content, could be potentially discruptive for media companies. People increasingly choose to read and watch content that either is reflecting their interests (yep, not everyone watches ESPN channel) or recommended by someone they know and trust. Also, opening your content to sharing (as opposed to locking it behind the paywall) leads to increase in new audience members, and engagement of current members. 

Collaborative Commerce: sharing and renting instead of buying, and loaning money from peers and crowdfunding your business and creative projects instead of turning to banks for credit could potentially turn to be discruptive for retail industry, manufactures, and banks. 

Mobility: business and personal productivity tools continue to move to mobile and virtual access-from-anywhere platforms, and that could potentially discrupt commercial real estate industry. Why rent overpriced office when I can rent a cubicle in a downtown loft for much less (electricity, Internet and coffee are included), or work from home or public park?

Gamification: more businesses are adding gaming elements (like badges, leader boards and rewards) to their business practices in order to increase customer loyalty and engagement. This could potentially lead to new ways of measuring key customer metrics and conduct market research.

Nielsen Releases Address-Based Sample TV Diary Data for Academic Analyses

This should be interesting:

As part of an effort to further research on address-based sampling (i.e., the sampling of addresses from a near universal database of household listings), The Nielsen Company is releasing data and paradata from its February 2011 Television Diary Measurement for analysis by eligible academic and student researchers. The February 2011 dataset contains information on respondents and non-respondents from 767,011 sampled addresses and includes more than 2,441 variables for each household record. The data/paradata capture all phases of recruitment and diary returns, including mail and call history information.  In addition, Census block group data and PRIZM data have been appended to each record.  Also, provided is a document that describes each variable—variable name, values, labels and origin/source for each variable.  The only data not included are the actual reported tuning & viewing information. The dataset is designed to allow researchers to examine in detail questions related to recruitment, nonresponse, mixed-mode approaches, etc. as they related to an ABS design.

The data are available free of charge to researchers and students at academic institutions; the data are not available to commercial or independent researchers. Users are asked to complete a Data User Agreement and agree to use the data for academic purposes only (i.e., courses, presentations, publications). No limits are placed on the types of analyses that may be performed. Nielsen does ask that researchers using the data to share with Nielsen all copies of reports, conference papers, manuscripts (etc.) that results from the use of this data and to note that the data cannot be used for commercial or proprietary use (see Legal Data User Agreement for details).

For further information, please contact: Anh Thu Burks, Manager, Research Methods – Center of Innovation,  Anh.Thu.Burks@Nielsen.com, 813-366-4368

 

AAPOR's 2012 Research Challenge

AAPOR just announced the 2012 Research Challenge:

To participate, researchers or research teams can submit a set of up to 10 survey questions for an upcoming in-person survey of Chicagoans. The “winning” submission(s) will be included in the in-person survey, and the winning researcher or team will receive the response data to their questions along with demographic data, and the first rights to publish results of their survey findings.

For full rules and contest information visit http://www.rti.org/researchchallenge 

Digging into Metro Card data

I had a very interesting discussion yesterday with a friend of mine who stopped using his Foursquare application, mainly because of the "privacy" concerns. He said that while he didn't mind getting all those discounts and special deals, he did not feel comfortable sharing his location with friends and acquantainces on Facebook and Foursquare, and a Big Brother. I suggested that he could opt for not sharing his particular "check-ins" with others by simply toggling a switch on his Foursquare app, and also pointed out that the reason that he is getting all those special offers and discounts is because venue owners and marketers expect him to share his visits to those venues, thus spreading the word. We then talked about how if he wanted to go "off the grid" completely, he would have to get rid of cell phones, stop using his credit cards, and switch from montly MetroCards to single ride passes (paid for by cash). This last point made me think about what kind of data is collected on metro cards and how these data could be possibly used by MTA, city planners and marketers. So I am started looking into that and can already share with you some of my findings so far:

Most of the data, including raw swipes and fare data at each turnstile are available on developers' site here 

This fascinating report by hackers' magazine 2600 on attempts to decode MetroCards data is certainly worth reading!

These awesome interactive graphs by Wall Street Journal dissect different aspects of Metro Card usage at Census block level, and and here is even more interesting blog post by the guy who actually analyzed the data and created these graphs

I just downloaded some of the turnstile usage data and will try to play with it later this week. Considering the available data points, these are the potential insights that could be extracted from the data:

- by looking at the turnstile data, one could calculate the average number of riders entering and leaving the station every 4 hours; it would also interesting to see if these numbers differ across the weekdays (e.g., more riders come on Mondays, etc.);

- by looking at the turnstile data, one could see how usage of different types of cards differ across the weeekdays and stations (e.g., if usage of certain type of carda peaks on certain day of the week; stations with the highest number of student or disability cards, etc.)

- by combining swipes and fare data with Census block data, one could see how usage is correlated with the neighborhood characteristics (e.g., Albert Sun looked at the correlation between the median HH income and usage of the unlimited 30-day metrocards).