Key Concepts Learned
- Algorithms: sets of instructions for solving problems or completing tasks
- Data analytics is inherently subjective
- Algorithmic decision-making in government:
- Assists or replaces human decision-makers
- Relies on models that process historical data
- Inputs: “features,” predictors, independent variables, x
- Outputs: “labels,” outcomes, dependent variables, y
- Promises efficiency, consistency, objectivity, and cost savings
- Margin of error (MOE):
- Large MOE relative to estimate = less reliable
- Small MOE relative to estimate = more reliable
- “M” in data tables stands for margin of error
Coding Techniques
- Clarified distinctions between:
- Data science: algorithms/methods, computer science focus
- Data analytics: applying data science methods across disciplines
- Machine learning: algorithms for classification and prediction that learn from data
- Artificial intelligence (AI): algorithms that adjust and improve across iterations
Questions & Challenges
- When will I figure out Github!?
Connections to Policy
- Real-World Examples
- Criminal justice/recidivism risk
- Housing & finance/mortgage lending
- Healthcare/patient care prioritization
- Government Data Collection History
- Civic registration systems, census data, administrative records, operations research
- Why Census Data Matters
- Foundation for understanding community demographics; allocating government resources; tracking neighborhood change; designing fair algorithms
Reflection
- I’m slowly getting things… I think?