Solution for Making Decision based on Large Set of Data

How Data Impacts Decision

Two Kinds of People in Decision Making Process

1. Liability: people who know materials well, but not involved in decision making process
2. Assets: people who can access materials and impact on decision making process

Ignore Data Scraps and Less Hunting

Need clear actions: Less Time Hunting Data, More Time Getting Actions

The Decision Making Modal

Stream Line

Decision

A choice or determination made after considering a set of presented materials

Supported Materials

1. Road Map before decision
2. Guidance for decision

Modals

1. Informed-ness
2. Timeless: decision-desired actions (time for taking action) => Risk Impact Analysis
3. Completeness: degree that I am confident in my decision
4. Preparation of Direct Action: streaming triggered actions
5. Optimality of Outcome: degree to which decision gets best result. This is the environment for best decision

The Decision Making Process

Step 1: Decision-Driven Decision Making

Always start from decision

What do we want to achieve?

1. What DATA can I access?
2. Problem-Solve-Targeted: We want to have our data available to subscribers.

Step 2: Find answers to questions

1. Know questions before see the data
2. i.e. What will answer my questions?
3. Draw three panels:

Step 3: Clean it up

Process in place before cleaning

Goal: Get to decision point

1. Build Process
2. Make data run under this process
3. Example: Get Excel File => Data Visualization => Find Rules to Clean File =>Get Cleaned File

How to clean data?

1. Capture: what bring attention? => Based on rules (error)
2. Clarify: what it mean in context (process) => looking for data connector points/common points
3. Organize: data with highest priority first
4. Reflect: review data frequently => Ask client questions. Note context may change
5. Engage: implementation with data

Step 4: Present

Target to the right stakeholder, right time, right way

Remember different people use data differently, so when present data to clients, remember to use data that client only cares.

Step 5: Distribute the insights

Apply insights to decision making process: Turn Data into Action

Three Panels for Evaluating Action

Examples and Conclusions

data.geogria.gov

Composition of Data

1. Based on decision I want to make
2. Bring historical data in
3. Ask questions: what cause the gap? What’s the impact of gap?

Tools Recommendation

1. CSVs
2. OpenRefine
3. SQL
4. DKAN
5. Data Integration List: share with client/standard for clean data. Example: null data? misspelling?
6. Majority of time spending on clean data => put this in SOW! show price in SOW!