Voronoi Diagrams

Voronoi Diagrams are ways to visualize the regions on a map that are the closest to a given node.

ggplot(spatial_results, aes(lat, lng)) +
  stat_voronoi(geom = "path") +
  geom_point(mapping = aes(lat, lng))

Old Method

Determining Market Size

We looked into the economic methods for evaluating market size. Two metrics we looked into were Reilly’s Law of Gravitation and Trade Area Capture. Reilly’s law is based on the idea people will tend to shop in areas with a higher population, and the model is based on a ratio between distance and population. Trade area capture uses historical data and income in areas to give a prediction of actual numbers of people who shop in an area.

Voronoi vs Reilly’s Law

Voronoi and Reilly’s law can tell us similar, but different things. Voronoi shows the geographic regions where our proposed store would be the closest. Reilly’s law would show the regions around our store that people would be willing to travel in order to shop at the store.

Aaron

This week Aaron worked on creating a function that takes latitude and longitude as inputs and outputs the corresponding Census geographies. So far, we have two: one that’s simple and one that is more complex. We are currently working on adjusting them and improving their performance, as well as figuring out how this function fits into the full patchwork of functions for our tool.

get_census_geos <- function(longitude, latitude) {
  
  address <- revgeo::revgeo(longitude = longitude, latitude = latitude, 
                            provider = 'google', output = "frame", 
                            API = Sys.getenv("PLACES_KEY"))
  
  city <- address$city
  
  state <- address$state
  
  state_ab <- state.abb[match(state,state.name)]

  
  # set options
  options(tigris_class = "sf") # use sf objects
  
  options(tigris_use_cache = TRUE) # cache downloaded files
  
  # get shapefiles for places in state
  places <- tigris::places(state = state_ab, cb = TRUE)
  
  # get shapefiles for counties in state
  counties <- tigris::counties(state = state_ab, cb = TRUE)
  
  # join places with counties with st_join
  city_county <- sf::st_join(places, counties, join = sf::st_within)
  
  # filter places by city name
  city_county_frame <- city_county |> dplyr::filter(NAME.x == city) |> dplyr::select(NAMELSAD.y)
  
  county <- city_county_frame$NAMELSAD.y[1]
  

  # county passed to TidyCensus
  
  census <- tidycensus::get_decennial(
    geography = "block", 
    variables = "P2_002N",
    county = county,
    state = state_ab,
    year = 2020,
    geometry = TRUE)
  
  
  census
}

knitr::kable(block_table[1:5,])

GEOID	NAME	variable	geometry
191279503001019	Block 1019, Block Group 1, Census Tract 9503, Marshall County, Iowa	P2_002N	POLYGON ((-93.029 42.02367,…
191279502002006	Block 2006, Block Group 2, Census Tract 9502, Marshall County, Iowa	P2_002N	POLYGON ((-93.11787 42.2077…
191279501002009	Block 2009, Block Group 2, Census Tract 9501, Marshall County, Iowa	P2_002N	POLYGON ((-93.04905 42.1297…
191279502002135	Block 2135, Block Group 2, Census Tract 9502, Marshall County, Iowa	P2_002N	POLYGON ((-93.13204 42.1058…
191279503004007	Block 4007, Block Group 4, Census Tract 9503, Marshall County, Iowa	P2_002N	POLYGON ((-92.98028 41.9495…

Srika

Harun

places_multiloc_pull <- function(df, name = NULL, 
                                 search_string = NULL, 
                                 keyword = NULL, 
                                 lat = c(df[,2]),
                                 lng = df[,3],
                                 radius = 6500, 
                                 type = NULL, 
                                 api_key = Sys.getenv("PLACES_KEY"), 
                                 place_type = NULL) {
  
  result_list <- list()
  
  for (i in 1:dim(test_cities)[1]) {
    
    start_time <- Sys.time()
    
    string <- sprintf("df_%f", i)
    
    spring <- noquote(string) 
    
    spring <- googleway::google_places(search_string = search_string, 
                                   name = name, 
                                   location = c(lat[i], lng[i]), 
                                   radius = radius, 
                                   keyword = keyword, 
                                   key = api_key)
    
    
    Sys.sleep(2)
    
    result_list <- append(result_list, list(spring$results))
    
    end_time <- Sys.time()
    total_time <- end_time - start_time
    print(sprintf("Execution time for iteration %d is %.2f", i, total_time))
  }
  
  return(result_list)
  
}

Conference

Miro

Meeting with Duane

Aaron

Duane Johnson, our client, stopped by East Hall Wednesday to answer key questions about our project. I had questions about the source of of some of the figures on the excel sheet that I had been working with, as well as questions regarding the extent our functions should take user input. After Duane answered our questions, the path has become much clearer regarding how to move forward with our project and what our final tool might look like.

One of my questions was about the depreciation sheet that was given to us. We asked whether or not we should provide default values for the assets listed in the calculation sheet in the case that the user did not know these values up front. Duane assured us that these can be purely user inputs and that we should not worry about default values.

Another question concerned the meaning of some of the line items on the sheet, such as “Operating Income (Loss)”. We were told that this represented an interim result in the calculation sheet and that this value is not to be forwarded to any other function. Rather, it is there to help us understand the progression of the calculation of pre-tax profit up until that point.

Finally, I asked about the final table on our recently updated excel sheet, which contained unseen columns such as “adv”, “repairs” and “tax”. Duane told us that this table represented the breakdown of the “Other Operating Expenses” line item on the sheet that we were familiar with. These are just a portion of the questions that were answered in regard to the Estimating Expenses excel sheet. I found our meeting to be highly productive and I have a much clearer idea of the big picture now.

Srika

Understood why the 2 methods were tried by the client to calculate the per capita grocery spending and planned to go with the simpler one moving forward. Got some clarity about the percent price increase and state index value used in Estimating revenue sheet. So we will use the consumer price index for the percent price increase and rural price parities as the state index as per our discussion with the client.

Alex

I discussed my findings of the week with Duane, and explained the Voronoi diagrams and Reilly’s law.

Next Week

Next week I plan on looking into how we can incorporate Voronoi diagrams into our leaflet map. The ggvoronoi package we are using does not save a geometry, so we will have to look into how we can map that. I also want to figure out if I can make a map using the Reilly’s law, because as is it will just give a single distance as an output.