2. Origins
• 2013-2015
o Legal Services Technology (LSC), Technology Initiatives
Grant (TIG)
o To develop data analysis technology strategies to better
serve clients
2
The Legal Aid
Society of
Cleveland
Montana
Legal Services
Association
Strategic
Data
Analytics
Northeast
Ohio Data
Collaborative
Cleveland
State
University
Northwest Justice
Project, LSNTAP
Strategic Data
Analytics
Scott Friday
Designs
• 2016-2017
o LSC TIG Grant, 2016-2017
o To create the Data Analysis Framework online tool to
help all legal aids use data strategically
5. Home: Things to think about
•Watch for data patterns
•Run every finding by staff
•Dealing with difficult data
•Data Integrity
•Factors that can skew your data
5
7. Link to detailed data
questions and
analyses
Analyses definitions
Recs re: internal and
external data &
academic partners
Link directly to
analyses that answer
data questions
7
1
2
3
4
8. 1. Data Questions
• High-level data questions
o Who is eligible?
= Poverty Population
o Who requests assistance?
= Intakes
o Who do we help?
= Served
o How do we help?
= Level of Service
o What resources are
required?
= Hours
8
• Click on a question box:
o Link to detailed questions
Link to analyses
9. 2. Analysis Types
• Snapshot
o Snapshot analyses measure counts or percentages for a given period, usually the most recently
completed year. If any counts or percentages are unexpected, comparison, trend or spatial analyses
may be necessary to better understand the reasons for the unexpected results.
• Comparison
o Comparison analyses review linkages between two or more variables and uncover information about
client conditions and data relationships. When unexpected data relationships are discovered,
investigation is warranted to better understand linkages and determine whether they indicate the
need for client service and advocacy work that simultaneously targets multiple conditions at once.
• Trend
o Trend analyses scrutinize changes over time in client conditions. Review trends over a five-year
period, or longer when possible. Spikes or dips that appear in trends might confirm what an
organization expects or raise additional questions worthy of investigation to better understand the
unexpected change and determine whether it calls for proactive steps.
• Geographic Distribution
o Geographic Distribution analyses show how people or problems or anything else of interest is
distributed across service areas, which can be divided into smaller areas to reveal spatial patterns.
These patterns are opportunities to learn about the spatial dimensions of your organization and your
clients.
• Geographic Concentration
o Geographic Concentration analyses compare geographic concentrations (high or low) of multiple
variables to determine how the variables and location impact each other.
9
10. 3. Data Resources: Internal
10
Case Data Fields: Client Data Fields:
Unique Case Identifier
Legal Problem Code
Open Date
Close Date
Case Status
Close Code
Outcome(s)
Poverty %
Persons Helped
Children in Household
Domestic Violence Involved
Unique Client Identifier
Race
Ethnicity
Gender
Age at Intake
Language
Education Level
Veteran Status
County and/or City
13. 3. Partnerships
13
University Departments with Data Analysis Capacity
State University Name Academic Department Department Website Law School Website, if applicable
Alabama ALABAMA A&MUNIVERSITY Biological and Environmental Studies http://www.aamu.edu/Academics/alns/bes/ESWSP/Pages/GIS-and-Remote-Sensing-Minor.aspx
Alabama ALABAMA A&MUNIVERSITY Department of Community & Regional Planning http://www.aamu.edu/academics/alns/crp/pages/default.aspx
Alabama ALABAMA STATE UNIVERSITY Department of History and Political Science http://www.alasu.edu/academics/colleges--departments/college-of-arts--sciences/history-political-science/minor-in-
Alabama AUBURN UNIVERSITY Architecture, Planning and Landscape Architecturehttp://cadc.auburn.edu/architecture/architecture-masters-degrees-programs/community-planning
Alabama AUBURN UNIVERSITY Department of Geology and Geography http://www.auburn.edu/academic/cosam/departments/geology/index.htm
Alabama AUBURN UNIVERSITY Department of Political Science http://www.cla.auburn.edu/polisci/
Alabama AUBURN UNIVERSITY AT MONTGOMERY Department of Political Science & Public Administrationhttp://sciences.aum.edu/departments/political-science-and-public-administration
Alabama JACKSONVILLE STATE UNIVERSITY College of Arts and Sciences http://www.jsu.edu/cas/
Alabama JACKSONVILLE STATE UNIVERSITY Geography/GIS http://www.jsu.edu/pes/geography/index.html
Alabama LAWSON STATE COMMUNITY COLLEGE Geographic Information Systems http://www.lawsonstate.edu/academics/careertech/gis/index.html
Alabama SAMFORD UNIVERSITY Department of Geography http://howard.samford.edu/geography/
Alabama UNIVERSITY OF ALABAMA AT BIRMINGHAM Department of Government http://www.uab.edu/cas/government/
Alabama TROY UNIVERSITY Department of Political Science http://trojan.troy.edu/artsandsciences/politicalscience/
Alabama UNIVERSITY OF ALABAMA Department of Geography http://geography.ua.edu/ http://www.law.ua.edu
Alabama UNIVERSITY OF NORTH ALABAMA Department of Geography http://www.una.edu/geography/
Alabama UNIVERSITY OF SOUTH ALABAMA Department of Earth Sciences http://www.usouthal.edu/earthsci/geo/index.html
Alaska UNIVERSITY OF ALASKA FAIRBANKS Department of Geography http://www.uaf.edu/snras/departments/geography/
Alaska UNIVERSITY OF ALASKA SOUTHEAST Geography & Environmental Studies http://www.uas.alaska.edu/arts_sciences/naturalsciences/geography/programs/index.html
Arizona ARIZONA STATE UNIVERSITY Department of Geography http://geoplan.asu.edu/ http://www.law.asu.edu
Arizona ARIZONA STATE UNIVERSITY School of Public Affairs http://spa.asu.edu http://www.law.asu.edu
Arizona NORTHERN ARIZONA UNIVERSITY Geography, Planning and Recreation http://nau.edu/sbs/gpr/
Arizona UNIVERSITY OF ARIZONA Landscape Architecture and Planning http://capla.arizona.edu/ http://www.law.arizona.edu
Arizona UNIVERSITY OF ARIZONA School of Geography and Development http://geography.arizona.edu/ http://www.law.arizona.edu
Arizona UNIVERSITY OF ARIZONA School of Government & Public Policy http://sgpp.arizona.edu http://www.law.arizona.edu
Arkansas ARKANSAS STATE UNIVERSITY Department of Criminology, Sociology, & Geographyhttp://www2.astate.edu/a/chss/departments/csg/
Arkansas ARKANSAS STATE UNIVERSITY Department of Political Science http://www.astate.edu/chss/polsci/
Arkansas UNIVERSITY OF ARKANSAS GIS http://libinfo.uark.edu/GIS/default.asp http://law.uark.edu
Arkansas UNIVERSITY OF ARKANSAS AT LITTLE ROCK Institute of Government http://ualr.edu/iog/ http://ualr.edu/law/
Arkansas UNIVERSITY OF ARKANSAS, FAYETTEVILLE Department of Geosciences http://geosciences.uark.edu/
Arkansas UNIVERSITY OF CENTRAL ARKANSAS Department of Geography http://www.uca.edu/geography/
17. Who is Eligible?
Snapshot Analysis
Example Analyses Steps:
1. Open the ACS, Advanced Search.
2. Click on the Geographies blue box on the left side of the
screen.
3. Select a geographic type from the drop down (in this
example: geographic type is state, state is Montana).
4. Click Add TO YOUR SELECTIONS and close the Select
Geographies window.
5. In the “topic or table name” box, enter B17024 or S1701
(depending on the data categories you need) and select GO.
6. From the list of tables that appear, click on the latest
available 5-year estimate.
•For information about choosing 5-year, 3-year, or 1-year estimates, click
here: http://census.gov/programs-surveys/acs/guidance/estimates.html
17
7. Download the table to Excel.
8. If the numbers downloaded into Excel as text, highlight the relevant cells, right click, and
select Convert to Number.
9. Perform calculations (including adding up all the numbers of people under 200% poverty
(because that is a good proxy for identifying all eligible people) from the various age groups
in the B17024 data).
10. Create a table like the one below in which the results of your calculations can be entered.
11. Create pie charts or other graphics, if helpful.
25. Example Analyses Steps:
1. Find the total number of intakes from your CMS for the last 5-10 years.
2. Create a table in Excel and enter the annual intake numbers in columns for each year.
3. Open the ACS, Advanced Search.
4. Click on the Geographies blue box on the left side of the screen.
5. Select a geographic type from the drop down based on the most appropriate type for your service area (state,
county, census tract, etc.).
6. Click Add TO YOUR SELECTIONS and close the Select Geographies.
7. In the “topic or table name” box, enter S1701
and select GO.
8. Download the S1701 table to Excel for your
area for the most recent 5 years. Note that if
your service area includes areas with
populations below 20,000, you should use the
5-year estimates.
• For information about choosing 5-year, 3-
year, or 1-year estimates, click here
9. Enter the numbers of eligible people from
each of the annual S1701 tables into the
columns for each year in the Excel file with the
intake numbers.
10. Create a combination chart in which intakes
are represented by a bar chart and the eligible
population in represented by a line chart on a
secondary axis
25
Who requests assistance?
Trend Analysis
Excel
26. Who requests?/Trend: MPBI Example
MPBI examples also here: 1. Who requests assistance? Snapshot --- 2. Who do we help? Snapshot
--- 3. Who do we help? Trend --- 4. How do we help? Snapshot --- 5. How do we help? Trend
26
27. Ex: Who do we help?
Geographic Concentration
27
29. Who do we help?
Geographic Concentration
Example Analyses Steps:
1. Export the total cases closed and served from your CMS to a spreadsheet for the most recently
completed year or the most recent year for which the ACS S1701 table is available.
2. Sort the served cases by county. Review the counties and remove any that aren’t actual county
names or aren’t in your service area. You may have to combine data if counties show up with
multiple spellings.
3. Subtotal all served cases. Then, calculate the percentage of served cases in each county.
4. Open the S1701 table and calculate the total poverty population for the state by adding up the Below
Poverty Level Estimate column for each county. Then calculate the share of the total poverty
population for each county. Add these percentages to a new column in your served cases
spreadsheet.
5. In a new column called Concentration, calculate the location quotient by dividing the served cases %
for each county by the % share of the poverty population and divide that amount by 100. Results
that are below 0.75 indicate that fewer clients were served than would be expected in that county
based on its share of the state’s poverty population. Results that are between 0.75-1.25 indicate that
the expected share of clients were served based on that county’s share of the state’s poverty
population. Results that are above 1.25 indicate that more clients were served than would be
expected in that county based on its share of the state’s poverty population.
6. Create a column called Concentration Ranges in which you enter these categories: “0.01-0.74”,
“0.75-1.25”, “1.25-3.00”, and “Less than 20 cases” (enter a threshold number of cases under which
you will not display the concentration data).
7. You should have a spreadsheet that simply shows County, Total Cases, Concentration, and
Concentration Ranges. 29
30. 30
Who do we help?
Geographic Concentration
8. Login to Microsoft Power BI
(create an account if you don’t
already have one).
9. Click on Get Data, then Excel,
find the spreadsheet you just
created, and click Open. Note
that your spreadsheet will need
to be in Microsoft Excel
Worksheet format for Microsoft
Power BI to import it into your
document.
10. Double click on the name of the
sheet in your spreadsheet and
then click Load.
11. Insert a Filled Map Visualization.
12. Enter County as Location and
Concentration Ranges as Legend.
13. Adjust the formatting as you
prefer to show the variation in
Concentration Ranges by county.
Make the counties with Fewer
than 20 Cases shaded white.
14. Use the automatic Legend or
create your own using shapes
with titles.
15. In order to include the map in
other documents, you will have
to take screen shots.
Interested in another
data question, click
Microsoft Power BI
33. How do we help?
Comparison Analysis
Legal Problem Code Race Brief Extended Grand Total
61 Federally Subsidized Housing African American (Not Hispanic) 66% 34% 100%
Hispanic 53% 47% 100%
White (Not Hispanic) 74% 26% 100%
Other 63% 37% 100%
61 Federally Subsidized Housing Total 65% 35% 100%
73 Food Stamps African American (Not Hispanic) 43% 57% 100%
Hispanic 40% 60% 100%
White (Not Hispanic) 63% 38% 100%
Other 74% 26% 100%
73 Food Stamps Total 45% 55% 100%
32 Divorce / Separation / Annulment African American (Not Hispanic) 86% 14% 100%
Hispanic 86% 14% 100%
White (Not Hispanic) 85% 15% 100%
Other 95% 5% 100%
32 Divorce / Separation / Annulment Total 87% 13% 100%
51 Medicaid African American (Not Hispanic) 52% 48% 100%
Hispanic 38% 62% 100%
White (Not Hispanic) 69% 31% 100%
Other 69% 31% 100%
51 Medicaid Total 51% 49% 100%
63 Private Landlord Tenant African American (Not Hispanic) 95% 5% 100%
White (Not Hispanic) 94% 6% 100%
Hispanic 92% 8% 100%
Other 93% 7% 100%
63 Private Landlord Tenant Total 94% 6% 100%
Example Analyses Steps:
1. Find the total cases closed with
both brief service and extended
service from your case
management system for the last
three years.
2. Using whichever analysis software
you prefer (Excel pivot table shown
in this example), sort data by legal
problems and limit your review to
the top 10 most prevalent legal
problems.
3. Further sort by Race.
4. Show percentage split between
brief and extended service.
5. Highlight results that deserve
special attention. In this example,
the data relevant to the questions
in the “Multiple analyses are
possible section” above are
highlighted in the table below.
33
Excel
36. Example Analyses Steps:
1. Export the total cases closed (including served or not served) from your CMS to a spreadsheet for the most recently
completed year.
2. Sort the cases by zip codes. Review the zip codes and remove any that aren’t actual five-digit zip codes. You may
have to combine data if zip codes show up in multiple ways (such as “87022” and “87022-“)
3. Subtotal the hours worked and number of cases by zip code. Then, calculate the average hours per case for each zip
code.
4. You should have a spreadsheet that simply shows Zip Codes, Total Hours, Total Cases, and Average Hours/Case. You
may want to add a column called “Country” that shows “United States of America” for every row in the spreadsheet
to help with geocoding later.
5. Login to Carto.com (create an account if you don’t already have one).
6. Go to Maps and click on New Map.
7. Click on Connect Dataset and Browse until you find the spreadsheet you just created. Click on Connect Dataset.
8. You may need to go into the Data View to change Zip Codes data from Number format to String format.
9. Still in Data View, click on the orange GEO box in the geometry column and select Postal Codes. Follow the steps to
enter the column name for Postal Codes (Zip Codes in this example) from the drop down menu of fields. For country,
either find the Country field in the drop down list or just type in “United States of America.”
10. Click on Georeference Your Data with Points or Georeference Your Data with Administrative Regions. When using Zip
Codes, select Administrative Regions to get the zip code boundaries to appear on the map.
11. Carto will geocode your data. When it’s done, click on Show. If the map doesn’t appear, click on Map View at the
top of the screen. Zoom in to see your service area.
12. Check out the interesting maps that Carto creates for you or edit the map any way you like.
36
What resources are required?
Geographic Distribution
37. Example Analyses Steps:
13. Click on the Map Layer Wizard and select Chloropleth. You may change the color ramp, size of markers, number or
buckets and other formatting.
14. If there are outliers (such as zip codes with just one or just a few cases), click on Filter and the select a column to
filter by, or click on the plus sign to add another filter. Select the Cases field and slide the left end of the chart so that
it only shows zip codes with 10 or more cases.
15. You can Add a Layer and go to the data library to find many built-in options, such as state, county or Census tract
boundaries.
16. You can also Change the Basemap to show highways, terrain, satellite images, and other options.
17. Save your map by giving it a new name and clicking Save.
18. In order to include the map in other documents, you will have to take screen shots.
37
What resources are required?
Geographic Distribution
Carto
38. Questions?
Rachel J. Perry
Strategic Data Analytics
Rachel.Perry@SDAstrategicdata.com
216-570-0715
Scott Friday
Scott Friday Designs
wsfriday@gmail.com
828-549-8286
Sarterus Rowe, Esq.
Northwest Justice Project, LSNTAP.org
sart.rowe@nwjustice.org
206-707-0811