Ghosts, Goblins, and Data? This Halloween, AMP & Predictify.me made 'BIG DATA'? Less Scary  

As a kid dressed as the latest Disney character costume or in classic scary attire, you knew which houses had the best candy and which houses were not worth lugging your pillow case down the driveway. In a way, it was your own scientific approach to Trick or Treating. But, imagine if there was a more scientific, proven method to finding the best candy and getting more of it. This Halloween, one trick-or-treater asked that exact question. His data scientist dad, Zeeshan-ul-hassan Usmani, Founder and CTO of predictify.me, with the help of AMP Agency, uncovered the answer. The mad Data Scientist approach consisted of the following steps: Collecting Data Candy Data Candy is small pieces of information that could be readily collected by kids while trick-or-treating their way through the neighborhood. Usmani's son wrote down the number and brand of candy that he got from each house during his three hour candy collecting spree. Data Candy Reporting & Analysis With the candy collected, it was time to sort the types of candy and input the information into a data scientists' most powerful tool ' a database in R.  In order to provide a more thorough analysis, we leveraged publicly available data sets including: Exact house addresses from Google Maps The worth of the house and monthly rent estimates from Zillow Names of the houses' residents, gender, age, political affiliation from public voters' database From which retailer the candy was purchased ' determined by geo-location information and an assessment of the retailers' online inventory What Did We Find? There is an optimal route for trick-or-treating. Out of 117 houses surveyed, 32 houses (27%) did not participate in Halloween. Usmani's son now has the data to lead his friends on the optimal trick-or-treating route in 2015. Lollipops, Twizzlers, Snickers and M&M's won Halloween. The top choice for Halloween treats among the surveyed group was lollipops (making up to 12.3% of all the treats), followed by Twizzlers, Snickers, M&M, Twix, Milky Way, Dove, Kit Kat, Whoppers and Tootsie Roll (with only 3.4%). Pay careful attention to who is giving you the candy. If the household consists of a woman between 41-to-50 years old and a man between 51-to-60 years old, there is 100% chance that you will get a treat. It dramatically reduces down to 25% if the woman is 51 years+ and the man is 60 years+. Similarly, if either of the household residents is 71 years+, you are less likely to get any treats.   Candy preference indicates political affiliation. This chart shows the preference of candy with the respective announced political affiliation of the people in the respective households. Blue represents that at least one resident in the house is Democrat (the other may or may not be UNA). Red represents that at least one resident is a Republican (again, the other may or may not be UNA). Lollipops and Kit Kats are highly associated with Democrats while Snickers and Butterfingers are associated with Republicans.   What does this mean for brands? Halloween is no joke. According to National Retail Federation,  Halloween is the second largest retail holiday behind Christmas: Americans spend $6.9 Billion on Halloween each year Americans spend $1.96 Billion on decorations Americans spend $2.6 Billion on costumes Americans spend $12.6 Billion per dollar per year on chocolate For brands, this type of treat-or-treat predictive analysis could help inform a more efficient, effective media spend during the holiday in addition a compelling Halloween messaging strategy for 2015. Interested in learning more about predictify.me and AMP's predictive analytics capabilities? Contact us or predicitify.me here. ### ABOUT PREDICTIFY.ME If the next five years looked like the last five, decisions would be easy.  But that's neither real nor accurate.  Predictify.me's proprietary HourGlass engine mashes massive amounts of data into a fully interactive simulation environment that allows you to test drive the future, with confidence

    Related Posts