Data Mining Policy and Practice

(The following are notes from an NCSL session about DATA MINING POLICY AND PRACTICE. These are my notes, as presented in the session, without my bias or input unless noted. I may or may not agree with what is said, but wanted you to know what lawmakers are discussing as they deal with DATA MINING POLICY AND PRACTICE).

** Data Mining exists in government sector, private sector, and hybrid situations

 

Steve DelBianco, Executive Dirtector, NetChoice

  • Expert on internet governance, online consumer protection, and internet taxation
  • Nationally known
  • Frequent Congressional hearings testimony
  • e-commerce expert

 What is data mining?

  • Discovering knowledge about data
  • Not tracking
  • Not real time analytics

Knowledge such as:

  • Patterns
  • Relationships
  • Tendencies
  • Sentiments

Better decisions can be made in areas of:

  • Business
  • Health care
  • Law enforcement
  • Government

Challenges to/with data mining:

  • Privacy
  • Civil rights
  • Consent of person

National Security Association (NSA) (Government/public sector)

  • Sources
    • Metadata from phone carriers
    • Tweets, posts, photos
    • Social media
  • Uses
    • Discover threatening content
    • Find network linkages
    • Pursue taps and investigate
  • Challenges
    • Obtain court approvals

 

Law Enforcement Data (Government/public sector)

  • Sources
    • Crime reports from multiple depts.
    • Conviction data
    • Demographic data
  • Uses
    • Crime predictor

 

License plate data (hybrid public/private sector)

  • Sources
    • Toll booths
    • Parking lot cameras
    • Secure facilities
    • On police cars and on streets
  • Uses
    • Ensure payment of tolls
    • Traffic and tourism studies
    • Learn when a car was near a crime
    • Challenges
      • Tolls purge data after a short while
      • Plate number, not personally identifiable

 

Retailer data on sales, returns and fraud (private sector)

  • Uses
    • Know what sells, when, to whom
    • Fraud score
    • Learn hi-risk shipping address of fraudulent transactions
    • Can identify “chronic returner”
    • They can then advertise on websites you visit

 

Social Media data mining (private sector)

  • Sources
    • Google, facebook, etc.
    • Friend networks created by users
  • Uses
    • Pattern and place of login
    • Suggest friends
    • Flag unusual login
    • Scan for illegal photos

Other areas of data mining and purpose:

  • Qualitative data
  • Tax Return data
  • MediCare fraud

 

Look at sources of mine, uses, challenges, and potential privacy issues

 

Doug Robinson, National Association of Chief Information Officers

What are states doing today? Top Ten state government uses of data:

  • Human services
  • Healthcare
  • Revenue and taxes
  • Finance, administration and procurement
  • Transportation
  • K-12 education
  • Labor and unemployment
  • Justice and law enforcement
  • Economic development
  • Higher education

Looking ahead….

  • Pressure to reduce costs in government
  • Broader adoption, maturity of analytics
  • Real-time dashboards
  • Challenges with state skill sets, competencies, recruiting most talented employees
  • Greater use of predictive analytics
  • Citizen input: sentiment about analytics
  • New tools, creative solutions

 

Christopher Soghoian, PhD, ACLU and Indiana University

  • Heavily published
  • Often quoted, referenced
  • Computer scientist by training, working within lawyers at ACLU (litigation, policy, surveillance issues)
  • After 9/11 we saw an incredible increase in illegal wiretapping (** ”illegal” is his word)
  • Companies may have good intention when designing these data mining tools, but eventually they make their way behind the closed door of the government
  • Government says in order to find the “needle in the haystack,” they need the information of the entire haystack
  • Government cannot promise us they will not get hacked, or the info will not be used by the next government administration
  • Why don’t we get more phone calls about privacy issues from constituents? They just don’t know it has been collected. (** um, yes they do, and they do call about it)
Share this post:

Leave a Reply

You must be logged in to post a comment.

Subscribe to our campaign e-mail updates!