Question 1

TO: State Director, National Organizing Team

RE: Weekly Metrics Report (7/10 - 7/17) & Past Goal Performance

In this week’s memo, I address individual percent-to-goals metrics for each organizer region, overall metric-based performance for the past couple weeks, identified issues with the current data, recommendations for improving the program, and suggestions for additional future data collection.

1. Weekly Goal Report, 7/10 - 7/16

I have decided to present individual percent-to-goal data for the week of 7/10 through 7/16. We have data for the period of 6/29 through 7/16, and without further direction on which specific week to analyze, I decided to present the most current seven-day period.

Below, you will find a table with each individual organizer’s campus/region code, their assigned weekly goal metrics, and there actual completed metric for each goal. From these numbers, I have calculated an additional percent-to-goal (PTG) metric for each goal this week.

Individual Percent To Goal Tracker
Goal Week of July 10-16, 2019
Campus Code Goal Metrics Completed Metrics PTG
Voter Registration Survey Cards Tabling/ Clipboarding Shifts One-on-ones Completed One-on-Ones Completed Survey Cards Completed Voter Registration PTG One-on-Ones PTG Survey Cards PTG Voter Registration
R1-A 10 18 6 2 0 27 9 0% 150% 90%
R1-B 4 11 3 2 0 1 1 0% 9% 25%
R1-C 3 8 2 2 0 6 4 0% 75% 133%
R1-D 5 10 3 2 0 14 5 0% 140% 100%
R2-E 13 24 6 2 0 19 17 0% 79% 131%
R2-F 10 16 6 2 0 16 12 0% 100% 120%
R2-G 6 12 3 2 0 8 5 0% 67% 83%
R3-H 10 19 4 2 0 34 6 0% 179% 60%
R3-I 10 22 5 2 0 2 1 0% 9% 10%
R3-J 6 15 4 2 0 14 3 0% 93% 50%
R4-K 10 18 6 2 0 16 9 0% 89% 90%
R4-L 9 20 4 2 1 49 12 50% 245% 133%
R4-M 3 12 3 2 0 11 1 0% 92% 33%
R5-N 13 20 6 2 0 18 5 0% 90% 38%
R5-O 5 20 3 2 0 2 0 0% 10% 0%
R5-P 6 12 3 2 0 16 10 0% 133% 167%
R5-Q 8 15 3 2 0 11 0 0% 73% 0%
Note: There is no 'PTG Tabling/ Clipboarding Shifts' column because unfortunately, no organizer has completed a shift in the given goal week July 10-17.
You will note the wide variability within organizers’ PTG for each metric, ranging from zero percent to significantly higher than 100% to goal. You will also note that no organizer has completed a single tabling or clipboarding shift for this week, and that only one organizer (R4-L) has completed a one-on-one this week.

That said, we can make a meaningfull assessment of individuals’ success based on their survey card and voter registration metrics. The following individuals are doing relatively well this week, hitting both metrics or significantly surpassing at least one of them: R1-A, R1-C, R1-D, R2-F, R3-H, R4-L, and R5-P. A special shout-out to R4-L, who has completed 245% of their survey cards goal!

Overall, four organizers hit or surpassed both survey card and voter registration goals this week, and an additional four more organizers hit or surpassed one of either goals this week. Therefore, eight of 19 organizers (42%) have met or surpassed at least one of their goals this week. While we should celebrate our success this week, the majority of organizers did not meet or surpass a single goal this week. Furthermore, it is worth repeating that only (1) one-on-one was completed by the entire team, and no tabling or clipboarding shifts were completed.

2. Historical Success Per Metric

The information above provides only a week of data, and we know that the success of our field program depends on months of work. Therefore, I am presenting the following information from our full time period of data (June 29 - July 16).

Specifically, which organizers are having the most success in specific goal areas? We will look at One-on-Ones, Completed Tabling/ Clipboarding Shifts, Voter Registrations, and Survey Cards Collected.

You may reference the four figures below to see each metric broken down by highest to lowest performing organizer in absolute terms. The top three highest performers, per metric, are:

One-on-Ones: R4-L, R1-A, (tie) R1-D / R3-H Tabling/ Clipboarding Shifts: R2-F (by a lot!), R3-H, R4-L Voter Registrations: R2-E, R4-L, (tie) R2-F / R5-N / R5-P Survey Cards Collected: R4-L, R3-H, R1-A

Let’s talk to R4-L, R1-A, and R3-H. They are performing at the highest levels in absolute terms across multiple metrics, and we should spread some of their best-practices to the rest of the team.

3. Issues with the Data

There are a number of identified issues with the data that we should correct on the front-end, to simplify data analysis. First, I identified a number of submitted dates that are not possible, including June 31, 2019, and February 29, 2019. June only has 30 days, and Feburary 2019 only had 28 days. Futhermore, there were multiple data formats in the sheet, which forced me to spend extra, unnecessary time standardizing the data. We should correct how we are collecting these time stamps. Second, there were a number of instances where recorded metrics were not assigned to a specific organizer, but rather their general region. Thus, we are unable to give credit to any particular organizer for these completed metrics. Finally, it is not clear when and where each goal week starts. I made an assumption in this case, but generally this is identified from the top.

4. Recommendations & Other Data

My primary recommendation is for the data and field team to identify why there is such variability between certain weeks in particular metrics, specifically tabling & clipboarding shifts and one-on-ones. Why was there such a focus on those metrics in weeks prior to the week of 7/10-7/16, but not during? And why did survey cards and voter registration increase so dramatically during this goal week? If our goal is to build a consistent field program that leverages all of our tools, organizers’ focus should be distributed proportionately across all four goals week to week.

Additionally, it would be helpful to assess the efficacy of individual tabling & clipboarding shifts or one-on-ones. How many voter registrations, survey cards, or volunteer shifts do organizers recruit during each shift or one-on-one? In the aggregate, this data could help us identify our most efficient methods for outreach, leading us to set better goals for organizers.

TO: Data Team

RE: Process

To construct my memo, I created an R markdown file in R Studio. I used a variety of library packages, including dplyr, tidyverse, gt, ggplot2, and ggthemes. Before importing all three sheets separately from the main Excel document, I changed the date formats to be standardized (this was when I realized there were a number of impossible dates recorded). To create the first section on one week of goal metrics, I filtered for the specific date range and “Completed” status before grouping by Campus Code and Category of outreach. Then, I summarized by the count of each category of outreach, and spread by type of outreach. I did this so I could join the completed metrics with the given weekly goals by individual organizer (Campus Code). With the unified data set between goals and completed metrics, I could create additional PTG columns and visualize the information in a table through the “gt” package.

For the second part of the memo, I removed the date filter, since I wanted to analyze aggregate performance per metric per organizer over the entire time period (6/29-7/16). Then, I created four separate bar charts per each goal metric, representing the total numbers per organizer in descending order. These were created using the “ggplot2” package.

Question 2

You are tasked with creating volunteer recruitment universes for your state. In addition to basic demographic information like name, age, contact information, scores, and campus, the following special tags and attributes are available: ● People we have registered to vote. ● People we have collected survey cards/ commit to vote cards from. ● Event completion history ○ Event types include 1-on-1s, rallies, meetings, canvassing, phone banks, text banks, tabling/ clipboarding on campus, and community events. ○ Event roles include volunteers, attendees, and admin work. ○ Event statuses post event include Complete, Resched, Cancelled, and No Show. ● Whether this person has expressed interest and checked “Volunteer” on their Survey Card ● The issues this person has checked on their Survey Card, including Healthcare, Climate Change, Cost of College, Cost of Rent, Equal Pay, Gender Equity, Gun Safety, Immigration, LGBTQ Equality, Racial Equity/ Justice. ● Whether this person has filled out an online form or taken another online action ● Whether this person opted to receive text messages from us ● People we registered or collected commit-to-vote cards from in 2018 ● Whether this person expressed interest in volunteering in 2017 and 2018

Using some subset of the fields above, describe how you would structure and build volunteer recruitment tiers. Include any tables or visualizations that would aid in explaining the design of your universe.

Question 3

  1. Using the Question 3 tab in the sample data workbook for reference, write SQL queries that return the following: (NOTE: I downloaded and renamed this sheet as ‘q3’ in my R Studio, and also edited the dates slightly before writing the SQL code.)
  1. The total number of unique people attempted by phone in the week starting 11/18/2019 by canvasser

SELECT canvasserid, COUNT(DISTINCT vanid) AS ‘Total Number of Unique People Attempted by Phone’

FROM q3

WHERE (datecanvassed BETWEEN ‘%2019-11-18%’ AND ‘%2019-11-24%’) AND contacttypename = “Phone”

GROUP BY canvasserid;

  1. The total number of phone attempts in that week by canvasser

SELECT SUM(*) AS ‘Total Number of Phone Attempts Per Canvasser’

FROM q3

WHERE (datecanvassed BETWEEN ‘%2019-11-18%’ AND ‘%2019-11-24%’) AND contacttypename = “Phone”

GROUP BY canvasserid;

  1. The result breakdown by different contact types

SELECT contacttypename, COUNT(resultshortname) AS ‘Result Breakdown by Contact Type’

FROM q3

GROUP BY contacttypename;

  1. The top ten canvassers by total number of successful contacts

SELECT canvasserid, COUNT(*) AS ‘Successful Contact’

FROM q3

WHERE resultshortname = “Canvassed”

GROUP BY canvasserid

ORDER BY canvasserid

LIMIT 10;