Data

Implementing AI Solutions in Business

Jacob H. Marquez

Data Scientist

Importance of data

No data, no patterns to learn
Wrong data, wrong patterns
AI is as good as the data

Big data for AI.

The four areas for data

$$ Acquire.

The four areas for data

$$ Volume

The four areas for data

$$ Quality.

The four areas for data

$$ Security.

Acquisition

Types of data

Customer demographics
Sales information
Product inventory
Marketing campaign performance

Supply Chain

Procurement
Transportation
Inventory levels

AI for images classification

Mechanical failures
1,000's of labeled images

Acquisition

First-party

Creative on data capture or existing sources
Speak to other groups
Take time to acquire

Third-party

Data source specialists
Partner with other businesses

$$ Volume.

$$

More complex = more data
Machine Learning scientist will determine requirements

Quality

A bad dish of food.

Quality

Completeness.

Quality

Accuracy.

Quality

Relevance.

Quality

Timeliness.

Quality - data bias

Definition: systemic errors or prejudices in the data that can lead to inaccurate or unfair outcomes

Tools and methods to recognize and evaluate
Prevent skewed outcomes and biased decision making
Build AI responsibly

Security

Helps to:

Maintain trust
Comply with regulations
Safeguard business interests

Measures:

Removing personal identifiers
Retention policies
Encryption
Access control

Let's practice!

Implementing AI Solutions in Business

Preparing Video For Download...