Capacitor Analysis

Author

Logan Scales

Introduction

A capacitor is a key component located inside the outdoor unit of a straight AC or heat pump system. It helps complete the circuit for both the fan and compressor. Because it is one of the most vulnerable parts of the system, it is also the leading cause of AC failures and often needs to be replaced. When the weather heats up and people turn on their AC for the first time, it is common for the system to fail to start due to a faulty capacitor.

In this analysis, I will be working with data from eBay. The buy-and-sell site represents a small but live sample market of HVAC parts that can give insights into pricing and demand. As an HVAC business owner, understanding what is being sold the most and at what price is crucial for the busy summer season.

Scope of Analysis

How can price and demand data for different capacitors from eBay help influence what to stock in summer months for my HVAC business?

This question of inquiry will be answered throughout the analysis by webscraping from eBay on the page where capacitors are being sold using methods from the rvest package. I will clean “messy” data in the form of data wrangling functions with the stringr package. Such as remove unwanted dollar signs or commas and add extra column for sizes or price groups.

Generating summary statistics such as the distributions and averages of prices will be a valuable base for this analysis. I will use bar graphs to view insights between price and units sold to see how price impacts demand. I will analyze pricing for the top-selling capacitor sizes and adjust service pricing to ensure the optimal price is given.

eBay provides a good idea of the market for these parts even if I’m not buying them directly from there. All data was sourced here: https://www.ebay.com/sch/i.html?_nkw=dual+run+capacitors&_sacat=0&_from=R40&_ipg=120

Analysis

Price Distribution

I will first examine the distribution of the price of capacitors with a histogram. This will give us a base to build off of as we work through the analysis. We will get a good understanding of the eBay market for capacitors that only shows a small sample but we will be able to tell the bigger picture of the market as a whole.

This histogram provides insights to how much we would be paying for each part when stocking up. The overall majority is right in the 10-15 dollar range. This demonstrates that the going price for the majority of the capacitors are fairly cheap. Stocking up and providing a competitive repair cost is possible. If these results showed a more even bi-modal distribution stocking and service pricing would have to be approached differently.

Sales by Price Group

Now that we know how much the majority of capacitors cost. I want to look deeper into if the majority are cheap does that mean those are the ones that are actually being purchased in largest quantities (represents demand). If I group the prices in 5 dollar segments then use a bar chart the express the average units sold by price group. The group that is getting sold the most can indicate market demand by what kind of AC people have. Each capacitor is identified by a certain size that goes with a certain size air conditioner, brands or price do not play a role in why one would be sold more than the other. Its a very practical market.

The result is pretty one sided showing that on average the most capacitors being sold are in the 20-25 dollar range. Already this first observation tells a different story than the first histogram we looked at. This also shows just because it is cheaper doesn’t mean it sells more it boils down to what you actually need and that price depends on what your previous version was. We can now move into trying to identify which size that is specially to know what to stock better.

Total Units Sold by Size

Now that we have established the price ranges that are getting purchased the most. I would like to identify the specific sizes that will be necessary to buy in bulk to have stocked up for when the repairs would occur. Capacitors are sized by a certain value then separated by either a “/” or “+” I have converted all sizing to be changed to a “/” for easy grouping. The first number (larger value) indicates the compressor size for the unit it is in and the second number (almost always a 5) indicates the fan size. The specific size capaictor that gets put inside an AC essentially is based on the size or internal wiring for that specific unit.

So according to a sample of sellers on eBay the most sold size is a 35/5, which makes sense this is a typical size for AC’s that are 2.5-3.5 Ton. That is a basic tonnage for most middle class homes. This is an indicator to make sure to have 35/5’s and 45/5’s stocked in the van for when the summer heat hits.

Average Price by Size

Now we know which capacitors to be stocking I want to plan for what prices to expect. The bar graph below shows the average price for capacitors by their size. So we can plan accordingly on how much we are going to have to spend on each capacitor size.

These results have the 35/7 being the most expense over 60 dollars on average. The capacitor sizes that we were interested in were 35/5 and 45/5. The 35/5 is around 40 dollars on average and the 45/5 is around 20 on average. As we work through this data further I’m making an assumption that the 35/5’s are so expensive ($40 is expensive for a 35/5) due to the eBay selling platform and that being the most popular by far and sellers have increased the price on that particular size.

Sales Distrbution by Price Group

Now I’m going to demonstrate a distribution via box-plots to express the units sold by the price group. This will give us an understanding overall on what we have been analyzing. This specifically goes beyond averages to show us how much sales can be vary for capacitors at different price points.

This box-plot visual reinforces what we have already learned. We already know that the largest subset right in the middle is mostly the 35/5. After knowing it’s price and popularity. The other really interesting thing we can pull from this data set is with the eBay market. You can almost see an equilibrium, if we look all the way to the left at the cheapest capacitors selling very little. As we move right the price increases but so do the sales. Once the price get more expensive after the middle the units sold decreases reverting back to the same pattern as we started. I think this is how the eBay marketplace has deveolped into when referring to capacitors. This visualization shows you the sweet spot if you will of the going price that will both get sales and most profit (if you happen to be selling capacitors on eBay).

Conclusion

If we were to piece everything together we now know which sizes we need to purchase a good idea on how many or little of the ones we need to buy. We also know around how much we are going to spend. If we were stocking our van with 30 capacitors at the beginning of this summer. I would feel comfortable with 10 at the size of 35/5’s which would be around $400. I would want 5 of the 45/5’s, 5 of the 40/5’s, and 5 of the 30/5’s totaling around $300. Then I would get 5 of a bunch of singles of different sizes averaging around $100. So in total around $800 for the starting batch of capacitors to start the summer.

With all that being said these numbers are produced from a very small sample on eBay where many HVAC companies are not buying their capacitors. The prices on eBay are slightly more expensive then if you would buy from an HVAC supply house. As for the purposes of this assignment I did not feel comfortable scraping product info from local supply houses because much of the scrape-able data was not even available unless I logged in with my contractor login.

If I were to do this analysis differently next time I would definitely reach out for permission so I can use some product data more tuned with the market or I would scrape even more data from eBay because I think value was still derived from the eBay data that is just not a place I would be purchasing my capacitors.

--- title: "Capacitor Analysis" author: "Logan Scales" format: html editor: visual toc: true # Generates an automatic table of contents. html: # Options related to HTML output. code-tools: TRUE # Allow the code tools option showing in the output. embed-resources: TRUE # Embeds all components into a single HTML file. execute: # Options related to the execution of code chunks. warning: FALSE # FALSE: Code chunk sarnings are hidden by default. message: FALSE # FALSE: Code chunk messages are hidden by default. echo: TRUE --- ## Introduction A capacitor is a key component located inside the outdoor unit of a straight AC or heat pump system. It helps complete the circuit for both the fan and compressor. Because it is one of the most vulnerable parts of the system, it is also the leading cause of AC failures and often needs to be replaced. When the weather heats up and people turn on their AC for the first time, it is common for the system to fail to start due to a faulty capacitor. In this analysis, I will be working with data from eBay. The buy-and-sell site represents a small but live sample market of HVAC parts that can give insights into pricing and demand. As an HVAC business owner, understanding what is being sold the most and at what price is crucial for the busy summer season. ### Scope of Analysis How can price and demand data for different capacitors from eBay help influence what to stock in summer months for my HVAC business? This question of inquiry will be answered throughout the analysis by webscraping from eBay on the page where capacitors are being sold using methods from the rvest package. I will clean "messy" data in the form of data wrangling functions with the stringr package. Such as remove unwanted dollar signs or commas and add extra column for sizes or price groups. Generating summary statistics such as the distributions and averages of prices will be a valuable base for this analysis. I will use bar graphs to view insights between price and units sold to see how price impacts demand. I will analyze pricing for the top-selling capacitor sizes and adjust service pricing to ensure the optimal price is given. eBay provides a good idea of the market for these parts even if I'm not buying them directly from there. All data was sourced here: <https://www.ebay.com/sch/i.html?_nkw=dual+run+capacitors&_sacat=0&_from=R40&_ipg=120> ```{r} #| label: Setup #| include: false library(tidyverse) # The tidyverse collection of packages library(httr) # Useful for web authentication library(rvest) # Useful tools for working with HTML and XML #Grabbing scraped data capacitor_df <- read_csv("https://myxavier-my.sharepoint.com/:x:/g/personal/scalesl_xavier_edu/EV0xWOdAWRdMoatKFK3HMTkBVWygQ5ICbxnfVAs46Qb4sQ?download=1") #Data Cleaning #Trims first 2 rows with nothing valuable in them capacitor_df <- capacitor_df[-c(1,2), ] #Removes extra unnecessary data in in the price column capacitor_df <- capacitor_df %>% mutate(product_price = product_price %>% str_extract("\\$[0-9]+\\.[0-9][0-9]") %>% #Logic sourced from ChatGPT str_replace_all("\\$", "") %>% as.numeric()) #Ensures that the only thing in "Units_sold" is numeric capacitor_df <- capacitor_df %>% mutate(product_units_sold = ifelse( str_detect(product_units_sold, "sold"), product_units_sold %>% str_remove_all("sold") %>% str_remove_all(",") %>% str_trim() %>% as.numeric(), NA )) #Removes extra junk from seller column capacitor_df <- capacitor_df %>% mutate(product_seller = product_seller %>% str_remove("\\s\\(.*")) #Logic sourced from ChatGPT was unsure #Only shows stars given capacitor_df <- capacitor_df %>% mutate(product_reviews = ifelse( str_detect(product_reviews, "out of 5 stars"), product_reviews %>% str_remove(" out of 5 stars\\.") %>% str_trim() , NA )) #Creates size column and converts sizing to all be seperated with "/" #Logic was soucred from ChatGPT capacitor_df <- capacitor_df %>% mutate(size = str_extract(product_name, "\\b\\d+\\s*[+/]\\s*\\d+(\\.\\d+)?\\b" )) %>% mutate(size = str_replace_all(size, "\\s*\\+\\s*", "/")) #Creates price groups capacitor_df <- capacitor_df %>% mutate(price_group = ifelse(product_price <= 5, "$0-5", ifelse(product_price <= 10, "$5-10", ifelse(product_price <= 15, "$10-15", ifelse(product_price <= 20, "$15-20", ifelse(product_price <= 25, "$20-25", ifelse(product_price <= 30, "$25-30", ifelse(product_price <= 35, "$30-35", ifelse(product_price <= 40, "$35-40", ifelse(product_price <= 45, "$40-45", ifelse(product_price <= 50, "$45-50", ifelse(product_price <= 60, "$50-60", "Over $60" )))))))))))) #I feel like there is a more effecient way of doing this. I was unsure. ``` ## Analysis ### Price Distribution I will first examine the distribution of the price of capacitors with a histogram. This will give us a base to build off of as we work through the analysis. We will get a good understanding of the eBay market for capacitors that only shows a small sample but we will be able to tell the bigger picture of the market as a whole. ```{r} #| label: Price Dist #| echo: false capacitor_df %>% filter(product_price <= 100) %>% ggplot(aes(x = product_price)) + geom_histogram(bins = 25) + labs(title = "Distribution of Capacitor Prices", x = "Price ($)", y = "Count") ``` This histogram provides insights to how much we would be paying for each part when stocking up. The overall majority is right in the 10-15 dollar range. This demonstrates that the going price for the majority of the capacitors are fairly cheap. Stocking up and providing a competitive repair cost is possible. If these results showed a more even bi-modal distribution stocking and service pricing would have to be approached differently. ### Sales by Price Group Now that we know how much the majority of capacitors cost. I want to look deeper into if the majority are cheap does that mean those are the ones that are actually being purchased in largest quantities (represents demand). If I group the prices in 5 dollar segments then use a bar chart the express the average units sold by price group. The group that is getting sold the most can indicate market demand by what kind of AC people have. Each capacitor is identified by a certain size that goes with a certain size air conditioner, brands or price do not play a role in why one would be sold more than the other. Its a very practical market. ```{r} #| label: Bar Chart #| echo: false capacitor_df %>% group_by(price_group) %>% summarize(avg_units_sold = mean(product_units_sold, na.rm = TRUE)) %>% ggplot(aes(x = price_group, y = avg_units_sold)) + geom_col(fill = "red") + labs(title = "Average Units Sold by Capacitor Price", x = "Price Range", y = "Average Units Sold") ``` The result is pretty one sided showing that on average the most capacitors being sold are in the 20-25 dollar range. Already this first observation tells a different story than the first histogram we looked at. This also shows just because it is cheaper doesn't mean it sells more it boils down to what you actually need and that price depends on what your previous version was. We can now move into trying to identify which size that is specially to know what to stock better. ### Total Units Sold by Size Now that we have established the price ranges that are getting purchased the most. I would like to identify the specific sizes that will be necessary to buy in bulk to have stocked up for when the repairs would occur. Capacitors are sized by a certain value then separated by either a "/" or "+" I have converted all sizing to be changed to a "/" for easy grouping. The first number (larger value) indicates the compressor size for the unit it is in and the second number (almost always a 5) indicates the fan size. The specific size capaictor that gets put inside an AC essentially is based on the size or internal wiring for that specific unit. ```{r} #| label: Sum of Units sold #| echo: false capacitor_df %>% group_by(size) %>% summarize(total_units_sold = sum(product_units_sold, na.rm = TRUE)) %>% filter(total_units_sold >= 500) %>% ggplot(aes(x = size, y = total_units_sold)) + geom_col(fill = "blue") + labs(title = "Total Units Sold by Capacitor Size", x = "Capacitor Size (MFD)", y = "Total Units Sold") ``` So according to a sample of sellers on eBay the most sold size is a 35/5, which makes sense this is a typical size for AC's that are 2.5-3.5 Ton. That is a basic tonnage for most middle class homes. This is an indicator to make sure to have 35/5's and 45/5's stocked in the van for when the summer heat hits. ### Average Price by Size Now we know which capacitors to be stocking I want to plan for what prices to expect. The bar graph below shows the average price for capacitors by their size. So we can plan accordingly on how much we are going to have to spend on each capacitor size. ```{r} #| label: Average price by size #| echo: false capacitor_df %>% group_by(size) %>% summarize(avg_price = mean(product_price, na.rm = TRUE)) %>% filter(avg_price >=20) %>% ggplot(aes(x = size, y = avg_price)) + geom_col(fill = "cyan3") + labs(title = "Average Price by Capacitor Size", x = "Size", y = "Average Price ($)") ``` These results have the 35/7 being the most expense over 60 dollars on average. The capacitor sizes that we were interested in were 35/5 and 45/5. The 35/5 is around 40 dollars on average and the 45/5 is around 20 on average. As we work through this data further I'm making an assumption that the 35/5's are so expensive (\$40 is expensive for a 35/5) due to the eBay selling platform and that being the most popular by far and sellers have increased the price on that particular size. ### Sales Distrbution by Price Group Now I'm going to demonstrate a distribution via box-plots to express the units sold by the price group. This will give us an understanding overall on what we have been analyzing. This specifically goes beyond averages to show us how much sales can be vary for capacitors at different price points. ```{r} #| label: Box Plot #| echo: false capacitor_df %>% filter(product_units_sold <=2500) %>% ggplot(aes(x = price_group, y = product_units_sold)) + geom_boxplot(fill = "lightblue") + labs(title = "Units Sold Distribution by Price Range", x = "Price Range", y = "Units Sold") ``` This box-plot visual reinforces what we have already learned. We already know that the largest subset right in the middle is mostly the 35/5. After knowing it's price and popularity. The other really interesting thing we can pull from this data set is with the eBay market. You can almost see an equilibrium, if we look all the way to the left at the cheapest capacitors selling very little. As we move right the price increases but so do the sales. Once the price get more expensive after the middle the units sold decreases reverting back to the same pattern as we started. I think this is how the eBay marketplace has deveolped into when referring to capacitors. This visualization shows you the sweet spot if you will of the going price that will both get sales and most profit (if you happen to be selling capacitors on eBay). ## Conclusion If we were to piece everything together we now know which sizes we need to purchase a good idea on how many or little of the ones we need to buy. We also know around how much we are going to spend. If we were stocking our van with 30 capacitors at the beginning of this summer. I would feel comfortable with 10 at the size of 35/5's which would be around \$400. I would want 5 of the 45/5's, 5 of the 40/5's, and 5 of the 30/5's totaling around \$300. Then I would get 5 of a bunch of singles of different sizes averaging around \$100. So in total around \$800 for the starting batch of capacitors to start the summer. With all that being said these numbers are produced from a very small sample on eBay where many HVAC companies are not buying their capacitors. The prices on eBay are slightly more expensive then if you would buy from an HVAC supply house. As for the purposes of this assignment I did not feel comfortable scraping product info from local supply houses because much of the scrape-able data was not even available unless I logged in with my contractor login. If I were to do this analysis differently next time I would definitely reach out for permission so I can use some product data more tuned with the market or I would scrape even more data from eBay because I think value was still derived from the eBay data that is just not a place I would be purchasing my capacitors.