The Importance Of Time In Analyzing Data Spikes

In March 2021, the Office of the Director of National Intelligence (ODNI) released a one-page executive summary of a coordinated intelligence assessment regarding the threat posed by Domestic Violent Extremists (DVEs). The trimmed-down assessment, which did not reference or include a data set, appeared to focus on “political and societal events” that occurred in 2020. Their analysis led them to forecast near certain violence by DVEs in 2021.

The ODNI did not release the full report. Analysts would have surely supported their findings with data. However, overall, they appeared to rely on a spike of activity by DVEs during 2020, or even into early 2021. A spike is a data point on a trend line that captures attention because it represents a change from the previous period. A spike can occur at any point along a trend line, but it is the recent upturn–a stark change from the previous period–that gets the most attention.  

Unfortunately, although new spikes draw notice, and sometimes alarm, they offer little opportunity for analysis because they lack the necessary element of time. Unfolding events cannot be properly analyzed, whether DVE activity, or other data sets, because intervening factors, such as investigation and prosecution, will temper the numbers. Some incidents may be reclassified, a prosecution may find a subject not guilty, or new facts may emerge that put the matter into a different perspective.

However, despite this reality, customers will be always be concerned about spikes and they will likely request an assessment. Here are a few suggestions to maximize the accuracy of a trend line analysis.

  • 1. Examine the trend within the context of a wide timespan—never one year–not even two or three–but 10 years or more. The longer the period studied, the higher the accuracy. [If a 10-year data set is not available, choose the longest period at hand.] 
  • 2. Compare/contrast. Analyze other spikes along the trend line. Have we seen this before? 
  • 3. Add perspective by examining/presenting the data quantitatively, such as the percentage change. 
  • 4. Consider several hypotheses that might explain the rise: did media attention, public outreach,  or a community awareness campaign drive new reporting?    
  • 5. Contextualize socio-political events. 
  • 6. Was there new legislation? 
  • 7. Did law enforcement put new reporting policies into place? 

As always, the idea is to disprove a hypothesis, not prove it.

Analysts may also be expected to forecast the direction the trend line will take. Will it continue to rise? Or will we see a drop? Although it may be expected, forecasting/prediction is largely futile and is not recommended.

Leave a Reply

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

You are commenting using your 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 )

Connecting to %s

%d bloggers like this: