2 of 17

Presentation Overview

Purpose

Provide an overview of Caltrans’ data literacy and analytics related training efforts

Summary

Caltrans overview
Data Literacy and Analytics Training Program Development
Leveraging Existing Expertise and Resources

3 of 17

California Department of Transportation

Headquarters (25+ Divisions)
12 District Offices
Approximately 19,500 employees (field and office staff)
Centralized Information Technology
29 Enterprise Data Stewards
12 District Enterprise Data Governance Liaisons

4 of 17

DATA LITERACY AND ANALYTICS TRAINING PROGRAM DEVELOPMENT

5 of 17

Why care about data literacy and analytics?

Provide an understanding to staff about why data governance and data management matter.
Ensure staff at all levels have the necessarily skills to create, comprehend, and/or work with data.
Support management desire for greater analytics capabilities.

Wisdom

Knowledge

Information

Data

6 of 17

Data Literacy Preliminary Investigation

The Preliminary Investigation gathered information in four areas:

Survey of practice.
Consultation with experts.
Existing data literacy programs and guidance.
Related research and resources.

Nevada DOT and Oregon DOT are currently developing training.
Minnesota DOT, Nebraska DOT and Sand Diego Association of Governments (SANDAG) has some training.
https://dot.ca.gov/-/media/dot-media/programs/research-innovation-system-information/documents/preliminary-investigations/pi-0285-a11y.pdf

7 of 17

Data Literacy and Analytics Training Program Contract

Develop a data literacy and analytics training program

Read and comprehend data
Conduct analytics, reporting and performance management activities utilizing business intelligence tools

Conduct a data literacy and/or data analytics capability maturity assessment and literature review to produce an as-is and to-be assessment and gap analysis

8 of 17

Training Topic Areas

Managing data
Understanding data
Data governance at Caltrans
Sharing Data
Documenting Data
Protecting and Preserving Data

Managing Data Quality
Improving Data Value
Being a good data citizen
Creating Data
Analyzing Data

9 of 17

Program Development Process

Literature Review and Existing Condition Assessment

Needs Assessment (as-is, to-be and gap identification)

Utilize capability/maturity model

Develop Training Approach

Student types (e.g., new hire, data analyst, executive)
Training topics by student type (tied to needs)
Training source (custom, off-the-shelf, hybrid)
Training delivery method (live instructor, recorded, print materials)
Course Catalog (outline, outcomes, cost, length, etc.)

10 of 17

Program Development Process

Develop Training Implementation Plan

Expected outcomes / benefits
Curriculum overview
Maintaining coursework and relevancy over time
Training communication materials
Performance tracking and reporting
Implementation resource requirements

Jump Start Training Plan Implementation

Marketing materials for executive management
Budget change proposal

11 of 17

LEVERAGING AVAILABLE EXPERTISE AND RESOURCES

12 of 17

Response to Statewide Needs:��CalDATA Data Literacy Share Out Series

State Chief Data Officer facilitated information sharing series
A different agency presented each week
Presentation held during lunch, recorded and posted online
Helps CA government agencies understand approaches and lessons learned and build peer network

Presenting Agencies:

Franchise Tax Board
City/County of San Francisco
State Water Boards

Health and Human Services
Social Services
Caltrans

13 of 17

Response to DOT needs:

Data Quality

Need: Support Business Areas
1Integrate Software
Workshops with 50+ Participants
Bi-Weekly Sessions with Homework
Develop Centers of Excellence
Takeaway: Importance of Data Literacy

DOT wanted to implement data quality management early on in data governance efforts based on staff feedback that data quality was a common issue across multiple lines of business and datasets. Staff did not want week long instructor led training so our Enterprise Data Manager developed a 16-session training program to cover the material with homework assignments in-between.

In bringing people together for the vendor training and in the subsequent internal training. It became obvious that many staff assigned with the responsibility over corporate data did not have a fundamental understanding of the concepts of how to manage and govern data beyond making non-relational databases and entering records into them. Staff did not understand terminology and concepts around data dictionaries, metadata, data quality management, as well as how to conceptualize business rules and apply them to the design of systems and QC/QA practices.

14 of 17

Data Analytics

Need: More Powerful Analytical Tools
Python Programming and Jupyter Notebook
Workshops with 30+ Participants
Bi-Weekly Sessions with Homework
Takeaway: Develop Centers of Excellence

Much the same as data quality needs, there emerged, a request for data analytics training to assist with BI / data visualization requests. The Enterprise Data Manager took a similar approach by taking free curriculum available from a university and began teaching the material to staff. Combined this with own expertise in using Jupyter Notebook and python to teach analytical techniques.

Lesson learned is that Caltrans needs to improve staff capability to perform analytics but we also developed the recommended approach of creating a data analytics center of excellence to address the most complicated analysis capabilities required to answer strategic questions and understand how and why things happen so we can enact changes that will be impactful and have the intended results.

Related to implementing data governance, the proposal is to create a permanent team of staff that act as subject matter experts that would work with individual business area data stewards and data custodians to jump start implementation of data governance practices and then provide on-call assistance to business areas. The CoE would also administer our data governance software covering data catalog, data quality, metadata management, and so on.

Related to data analytics, the proposal is to create data analytics teams within different programs that would be a resource to multiple business areas within each program (e.g. Project Delivery, Maintenance and Operations, Planning and Modal Programs). These permanent teams would answer the most complex questions and also provide support to report creators in each business area as needed. As part of this work, the CoE may recommend changes such as what data is collected or the structure of data in order to be able to perform the required analysis. The CoE answers the “how” and “why” questions while the business area staff still handle the “what” items. This could change somewhat over time as business area staff improve their capabilities.

15 of 17

Data Gov. Task Force

Need: Support Data Visualization and Publication
Cross-Functional Team on 12-Month Term
15 HQ Divisions participating across 17 teams
Support Business Areas in Data Governance
Create data visualizations and GIS web maps
Takeaway: Importance of Change Management

Caltrans implemented a 12-month task force to help business areas implement data governance practices. This included on the job training to understand the basic concepts of documenting data, managing quality and creating data visualizations. The idea was to jump start activities within multiple business areas and teach staff the basics and make them change agents within their own business areas.

Task Force Participant Business Areas Representing 12 Divisions:

Accounting and Budgets, Transportation Planning, Maintenance, Project Delivery, Engineering Services, Financial Programming, Local Assistance, Rail and Mass Transit, Environmental Analysis, Strategic Performance, Equity, Safety

Benefits:

1. A larger enterprise data governance team.

2. EDGTF will expedite Caltrans’ data governance implementation.

3. Efficient use of limited consulting dollars available.

4. Increased visibility for enterprise data governance at data custodian level.

5. Data consistencies and improved communications across systems.

Lesson learned: There is a broad lack of understanding of the concepts of data governance across business areas. People in charge of data haven’t necessarily been trained in and follow best practices nor do they fully understand the importance and steps to implement these practices. Staff say they want quality, accessible data but don’t fully grasp what that means or how the data governance efforts will assist. Data literacy training curriculum will help with this as well as broad data governance change management activities we plan to undertake..

16 of 17

Outreach and e-Learning

Need: Workshops and Knowledge Sharing
American Stat. Assoc. Data Challenge (Mar 2021)

Project: GitHub Repo

CA Water Board Trash Datathon (Feb 2021)

Recording: YouTube
Project: Google Colab

Code for America Equity Asset Map (Sept 2020)

Event: CFA Website
Project: Google Colab

CA Stormwater Conference (Sept 2020)

Event: CASQA Website

LinkedIn Learning

Website: Data Governance

Takeaway: Data Community Participation

We recognized and leverage external opportunities to reach out to peers and practice skills through various engagements with sister agencies, non-profit organizations, etc.

We also looked at what related training we could find within the training programs already being paid for.

Lesson learned is that the outreach activities not only improved our skills but strengthened relationships between agencies that goes beyond the topic of data literacy and analytics. There is considerable data governance/management related work going on outside of the transportation sector that is directly applicable to our challenges. Building this greater awareness and relationships has been beneficial in that we have a growing set of resources to call upon.

We also were able to leverage these connections to obtain feedback to our approach regarding open data and increasing our transparency by reaching out to groups such as Code for America and asking them to review our data offerings and provide feedback on value and usability.