Data Analyst with strong problem-solving skills. Extremely organized and dependable person with advanced capabilities in Excel, SQL, Tableau, and Python.
Intro
Hello, I’m Andrew Brock. I am a data analyst with strong problem-solving skills. I am a very organized and dependable person that is comfortable working with groups or on my own.
My experience in teaching math allows me to deliver complicated information in an easy-to-understand manner. I’m excited to bring my unique background and analytical skills to a team to contribute to business decisions and growth of the company.
As a former math teacher, I have developed exceptional analytical skills and a sharp eye for detail. I enjoy working on complex problems and translating raw data into actionable insights.
With a deep understanding of statistical methods, SQL, and data visualization techniques, I am skillful at identifying patterns and trends, and providing strategic recommendations.
I am committed to continuous learning in the dynamic field of data analytics. I am eager to showcase my skills and provide a fresh outlook to an organization. Please explore my website as it displays a few of my projects I have completed.
resume
projects
Instacart Grocery Basket
Background
Instacart is an online gorcery ordering and pickup service that is looking to unccover more information on their sales data and customer puchasing behaviors.
Key Questions
1. What customer profiles make up the majority of sales?
2. Which products are the more popular?
3. What hours are the most/least popular?
4. Which products are in large supply?
Goal
Use the given data sets to provide data-driven insights on customer profiles and spending behavior to help improve Instacart's marketing strategies.
To get a good understanding of the data, I imported necessary libraries and datasets into a Jupyter Notebook. Used data wrangling and cleaning techniques to make sure the data was acceaptable to use for analysis.
Merged all necessary data, created new variables and exclusion flags. Used comments within the code to make it accessible to others reading my analysis.
Data Analysis
Instacart should consider a marketing campaign around Personal Care, Household, and Pantry items as they are in large supply, but not many orders.
Determine when orders are being placed. We can see that the bulk of the orders are happening on Saturday (0) and Sunday (1), and the most popular shopping hours are from 9am - 4pm.
Instacart has many family profiles that use their service, but he largest customer base are those that are married with children.
Recommendations
1. Target customers that are married with children as they occupy the largest family profile group.
2. For the marketing campaign, use popular products such as beverages, dairy, and produce. Also consider personal care and household items because of the high supply.
3. Schedule advertisements during the weekdays and outside of the hours of 9am - 4pm as these are the periods with the fewest orders.
Next Steps
Closely monitor sales after running appropriate advertisements to the targeted profile group and for the suggested products. Get additional feedback from family groups to see if the advertisements are impacting buying decisions.
Rockbuster Stealth
Rockbuster Stealth Video Rental company
Coming Soon!
Olist Brazilian E-Commerce
Background
The Brazilian E-Commerce company, Olist, is an online company that brings together individual sellers and buyers from all over Brazil in one, easy-to-use, place.
Key Questions
1. Are there any common factors that influence a review score?
2. Where are most of the orders coming from in Brazil?
3. What product categories are the most popular?
4. What categories would be beneficial for new sellers to break in to?
Goal
The goal of the analysis is to gather insights about Olist to see how the company can improve in a way that increases revenue and makes the customers happy.
Imported all of the necessary libraries and datasets into a Jupyter Notebook. I was able to use common variables to merge the data into one useful dataset. From there I could apply cleaning methods, exploration, and descriptive statistics.
Missing values, duplicates, inconsistant data types, unnecessary columns.
Data Analysis
Consider correlations, time series analysis, linear regressions throughout analysis with Python. Use Brazil States .json file to make a Chlorpleth map to gain a better understanding of the E-Commerce layout.
Import data into Tableau to explore other variable relationships. Determine the reasons for poor review score as well as good review scores.
What items are selling the best and where are people purchasing these items?
From this graph we can see a drop in average review score during the time period of December 2017 - March 2018.
Comparing to the graph below, the top 10 products had a large spike in total revenue, as well as slower shipping times during the same time period.
Each order has an estimated deliever date. An order that is considered 'delayed' would be an order that arrived after the estimated date.
There is a clear difference between orders that arrive on time and orders that are delayed for their review scores.
Recommendations
Orders that are arriving 'on-time' are recieving a review score that is 63% higher than the review score of an order that was delayed.
1. Focus on the high revenue products such as the 'health and beauty' and 'bed, bath, and table' categories.
2. Expand further north by adding additional sellers in those areas.
3. Determine what causes orders to be delayed to prevent it from happening again in the future.
4. New sellers would benefit from a high demand and low supply of sellers in areas such as 'wathes and gifts' and 'bed, bath, and table'.
Next Steps
Assess what occured during the time period that had an increase in shipping delays to hopefully prevent it from happening again.
Focus on the items that sell, and continue to increase the number of sellers in those areas.
Continue to monitor and address any changes in shipping or selling patterns to avoid a drop in review ratings.
Influenza Season Preparation
Background
During the Influenza season, hospitals become overcrowded with patients which leads to an increase in deaths. Staffing agencies need to determine where to send additional staff to lower the death rates.
Key Questions
1. What states are in need for the additional staff?
2. When does the additional resources need to be deployed?
3. Who is the flu effecting the most?
4. How efficient are vaccines in slowing down the death rate?
Goal
The goal of this project is to use prior knowledge and data of Influenza seasons to determine where the additional staff is needed and when they should be deployed.
Conduct descriptive statistics and visuals to get an idea of the data. Profile data and address and inconsistencies.
Use pivot tables with Excel for quick calculations. Test hypothesis, find correlations, and aggregate the necessary data.
Data Analysis
The vulnerable population is defined as indivuduals that are 65 years or older.
When comparing the number of vulnerable deaths to the total number of deaths, we can see that in the areas where there are not as many total deaths, the vulnerable deaths make up a larger percentage.
The Influenza season flucuates between states, but on average, it starts around November/December and goes until March/April.
In preperation for the 2018 Influenza season, we will focus on the years 2015-2017 to best determine where to send the additional resources. We can determine that Hawaii has the greatest need for additional resources. Follwed by Mississippi, Tennessee, and New York.
1. Additional staff should be provided from November to April.
2. Influenza vaccines are highly encourged for those 65 and older.
3. Prioritize states with a high death rate such as: Hawaii, Mississippi, Tennessee, New York, and Missouri.
Next Steps
After the 2018 Influenza season, we can compare how the states improved on their death rate for the vulnerable population, and assess any new states where the death rate increased. Did we divide the resources appropriately and to the correct states. Also look further into vaccination status and how big of an impact it plays for preventing deaths.
Career highlights
As an educator with 10 years experience of taking complex ideas and presenting them in a simple and concise way, I have found a new joy in analytics. I incorporate my problem solving and technical skills of Excel, SQL, Python, and Tableau during my current role as much as possible.
I have learned a lot from my experience in the CareerFoundry Data Analytics Course. With a combination of my drive and problem solvings skills, as well as the techniques I have practiced in this course, I believe I would instantly become an asset to a team that is looking to succeed.
Problem solving
As a teacher, there are many additional responsibilities that go on at school. Our students would vote for various things throughout the year and the Administration would always have to sort and count the votes by hand. In a small school, this wouldn't be a big deal, but in our school with over 1,200 students, this could take a while.
I went to the Administration to see why they didn't want to do something like an online form. They expressed valid concerns of students voting twice, students giving out the link so they could recieve extra votes from family or friends from other schools, and also students voting outside of their grade level.
I considered their concerns and knew I could solve all of those issues with SQL. I first set up a Microsoft Form for students to vote for Homecoming. We sent it out to all of the students telling them to just vote within your grade level and only vote one time. When the voting was over, we could initially see the votes within the Microsoft Form page, but I knew this data could be inaccurate.
I downloaded the results as a .csv file and create a table in PostgreSQL to input the values. I joined the raw results with our student_accounts data ON the student email address. This would null any vote that was not from a student from our school. Then, I grouped distinct results by grade level and I was able to see the number of votes for each grade level for each participant.
I quickly sent the results to Administration and they were able to communicate the results to the students in minutes with confidence and without mistakes.
I saved Administration hours of time and now we use the queries for every vote we have in the school.
Communication
Constant communication with students, coworkers, and parents to increase awareness of students' abilities and understanding of grades. Consistent communication with all individuals involved leads to better understanting of expectations.
Technical
Cleaned, filtered, and sorted SAT score data for entire school. Assigned students to groups according to teacher and score for each topic. This allowed for individualized instruction for each student.
Utilized excel for attendance, grades, and master schedule of students. Used conditional formatting and created formulas to reduce class sizes and prevent overlap of students.
Results
Increased class average by 38% on the SAT by identifying strengths and weaknesses through data driven teaching.
My technical, as well as my problem solving skills, allowed for us to identify strenths and weaknesses of various groups to specifically target their individual needs. This process had a direct impact on SAT scores.
previous employers remarks
"Andrew does a great job in the classroom varying his teaching strategies and differentiating instruction. He incorporates technology and interesting activities, while focusing on math skills. His students respect him and are engaged in the learning process." - Dr. Branch, Principal.
This is bold and this is strong. This is italic and this is emphasized.
This is superscript text and this is subscript text.
This is underlined and this is code: for (;;) { ... }. Finally, this is a link.
Heading Level 2
Heading Level 3
Heading Level 4
Heading Level 5
Heading Level 6
Blockquote
Fringilla nisl. Donec accumsan interdum nisi, quis tincidunt felis sagittis eget tempus euismod. Vestibulum ante ipsum primis in faucibus vestibulum. Blandit adipiscing eu felis iaculis volutpat ac adipiscing accumsan faucibus. Vestibulum ante ipsum primis in faucibus lorem ipsum dolor sit amet nullam adipiscing eu felis.
Preformatted
i = 0;
while (!deck.isInOrder()) {
print 'Iteration ' + i;
deck.shuffle();
i++;
}
print 'It took ' + i + ' iterations to sort the deck.';
