Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit

ITEC-320

Google Trends Activity – Identifying Fads

Google Trends provides data on the relative popularity of a search term over time.  Search term popularities exhibit different patterns; some remain fairly steady over time, and others are very high for only a short period of time.  Marketers refer to the latter as fads, and are interested in understanding them more thoroughly for many reasons.

Because the number of possible search terms is astronomically high, it would not be practical for humans to identify all current and recent fads manually.  The goal of this activity is to develop a metric that uses Google Trends weekly data for a search term over the past five years to measure the extent to which that search term is a fad.

Use the following steps to get started:

1. Open a browser and go to Google Trends (https://trends.google.com/trends/?geo=USLinks to an external site.
).

2. Enter the search term “The Office.” You should then be shown a page with a chart showing the “Interest over time” for searches on “The Office.”  Notice that it is fairly steady over time.

3. Above the chart, you should see several options that can be changed. Change “Past 12 months” to “Past 5 years.”  The chart should update to show the previous five years of search data.

4. Click the download icon on the right above the chart to download a CSV file containing the search interest data. Open up that file.

5. Column B contains the popularity level for each month. Notice that the highest value is 100.  Regardless of what search term you’re looking at, the highest popularity level will always be set to 100.  All other values can be interpreted as percentages of that highest level.

6. In your browser, change the search term from “The Office” to “Tiger King.”

7. You will probably notice that this chart looks very different. This search term could be considered a fad.  Download this dataset as a CSV as well, and open it up in Excel.

8. You may have to do some data formatting, since there will be some “<1” values that will be treated as text.

9. Again, the highest value in Column B is 100. However, you should notice that the values in Column B behave very differently in this dataset.

10. Try to express clearly, quantitatively, exactly how this dataset differs from the previous one.

The challenge is now to create a metric that can be used to determine whether or not a column of Google Trends data represents a fad.  The metric should be computed using a single Excel formula entered into cell B265, and it should refer to the data in B4:B264.  It can either be binary (Yes/No), or it can be a number that is higher for fads than for non-fads.  There are many possible metrics that will work.  Experiment until you are confident that your metric can distinguish fads from non-fads.  You can try downloading data for other search terms from Google Trends if you would like to test your formula on new data.

Once you have settled on a formula for a fad metric, go to the Google Trends Activity on Canvas.  On the submission page, upload a Word file that contains the following:

1. The full names of each of your group members.

2. The Excel formula used to compute your metric.
The formula itself is what needs to be submitted, not the result. If your formula is “=STDEV.S(B4:B264)” then that is what should be in your document, not the resulting number.  (Do not use that specific formula; it is not a good one for measuring fads.)

3. A short (150 words maximum) justification for your metric.
Convince your instructor that it is a reasonable way to detect fads. You do not have to argue that it’s the best possible metric; just explain the reasoning behind it and why it will work the way it’s supposed to.

Only one submission per group is needed.

50% of your grade will be awarded for a proper and complete submission before the end of class, including a coherent justification.  For the other 50%, your instructor will apply your formula to the interest levels over time for five pre-selected search terms.  If it works correctly (i.e. all of the Yes/No labels are correct if binary, or all of the fads receive higher scores than the non-fads if numeric), you will receive full credit for this 50%, otherwise 10% will be deducted for each misclassified search term.

You can certainly use an IF/Then statement or Nested IF/Then statement which will return the terms "Fad" or "non-Fad" based on the criteria you have chosen.

Alternatively - your formula can just return a Numeric value.  You must then tell me what numeric value represents in your submission. 

For example - you could say, "Search terms with a value of 50 or greater are Fads, and 49 or less are non-Fads".

This is the condition you have defined for your formula being able to distinguish between the two. 

Obviously - if your formula has the terms as part of an If/then statement - you don't need to explain that.  However, you should explain your approach and why you chose it (part 3 of the assignment).