Industry sign banking north dakota job description template computer
my name is Ari lamps team and in this presentation I'm going to give you an introduction to the choropleth er suite of our packages this presentation will tell you everything you need in order to get started with all of the main features of choropleth er it will also point you to resources where you can learn more about specialized aspects of choropleth er each of the maps you see here is called a choropleth map a choropleth is any map that shows regions for example US states counties or countries of the world and express his values for those regions through color each of the maps you see here was created with choropleth err let's start with some common questions that I get the most common question I get is why did I name the package choropleth er a lot of people have trouble pronouncing the name the name comes from combining the words choropleth map with our programming language another common question I get is why did I create choropleth er there are already a number of our packages that create Maps or that connect to open datasets the answer is that I developed choropleth or while working at an online real estate company and I needed to visualize all of the Sales Lead data my job was to help optimize how we distributed leads and I ran a lot of sequel queries that I needed to quickly visualize I couldn't find in our package that did exactly what I needed and so I created it myself as time went on I became more and more interested in visualizing demographic statistics and so I extended choropleth to connect to a lot of AP is for example the US Census API and that allowed me to import analyze and visualize that data and to the best of my knowledge there's no other r package that quite does that there are a few things that might help you learn choropleth er if you know right up front the first is a choropleth ur was designed specifically for exploratory data analysis as you'll learn soon this means that choropleth R values allowing you to look at a particular data set in several different ways very quickly choropleth ur is a large project and as a software engineer I tend to make sense of large projects by enforcing strict rules and naming conventions and so while choropleth ur is quite large I think you'll find it easy to navigate if you learn these few rules sometimes people want to know a bit more about me my name is Ari Lam steam and my website is re Lam sting comm I'm the author of choropleth ur these days I work as an art trainer and consultant and before this I worked as a software engineer and data analyst at various tech startups in San Francisco here's the outline for my talk we're about halfway through the introduction after this slide I'm going to give you some quick wins with choropleth ur then we're going to go behind the scenes and look at what choropleth er does when you issue these simple commands once you've mastered that material we're going to look at all the maps that core plethora ships with then we're going to look at all the data integrations that core better has choropleth ER connects to the US Census API the World Bank API then it packages some data from the Bureau of Labor Statistics as well and finally I'm going to briefly give you some pointers if you want to learn more about how to use choropleth er this is the map that most people start with when they learn choropleth er first of course you have to install the choropleth or package and then you need to load it with the library command choropleth ER comes with a lot of example data here we have a data frame called DF pop state and you can learn more about it by typing question mark DF pop state it contains US state population estimates we can create a map of that data frame by typing state choropleth DF pop state and I also encourage you to read the help for the state quarter plus function by typing question mark state choropleth one thing I like to point out is that there is never just one right way to visualize a dataset if we look at this data set you can see that there are seven quantiles which means there are seven discrete colors and if you look closely you'll see that the data set ranges from 560 mm to 37 million there are some patterns that really stand out for example these four states are right next to each other and seem to have very low populations also while California Texas and Florida are all dark blue it's not clear if they all have the same population or a different population if you set the number of colors to one by typing num colors equals one kora pother will use a continuous scale and this makes it very easy to see outliers in the data for example on this map it becomes very clear that California has the largest population of all states in America larger even than texas or Florida finally choropleth ER lets us zoom in to specific states we do this by setting the parameter zoom equal to vector of states that we want to zoom in on and here we have a map that zoomed in on California Oregon and Washington finally choropleth er lets us superimpose a choropleth map over a map from Google Maps we do this by setting the reference map equal to true and here for example we can see where large cities are located for example San Francisco is in the northern part of California while Los Angeles is in the southern part and while this doesn't relate to explore to our data analysis choropleth are also lets you add a title and legend to your map by setting the title and legend parameters here's a brief summary of what we've learned so far Quorra plethora lets you create a map of US states by simply calling the function state choropleth it also gives you a number of options to help you explore the data you can change the scale with the num colors parameter you can zoom with the zoom parameter and you can add a reference map with the reference map parameter although I should mention that Google Maps recently made a change to their API and in order to use the reference map parameter you now need to get a free API key from Google Maps now that we've learned the basics of choropleth ur we can go behind the scenes and see how it works while I've been a software engineer for many years this is actually the first time that I created or saw a single function that lets you create a map with a single piece of data and so it's a bit tongue-in-cheek here to say a state choropleth a magical function but it certainly is a bit unusual the best way to learn what choropleth err is doing is to look at the data set DF Pop State and you can see here that it's a data frame that has two columns one column is named region and the other column is named value and this is a rule that we'll run into over and over again with choropleth er your data set needs to have one column named region and one column named value next I want to point out that the map that choropleth er uses is something that you can see yourself you will need to load a new library choropleth ER maps but once you do you can see that the map is called state map and if you type question mark state map you can see some information about it if you look at the map you'll see in particular that there's one column called region and this is how choropleth are works when you call the function state choropleth with a data frame that data frame needs to have a column named region and a column name value behind the scenes choropleth ER merges your dataset with state map using the column region and this is why your regions need to be spelled exactly the same way they appear in state dot map finally colors are assigned based on the value column and the maps are rendered with ggplot2 if you read the help file for State Quarter Plath you'll see a description of data frame D F this is the user supplied data the description is a data frame with a column named region and a column name value elements in the region column must exactly match how regions are named in the region column in state map I like to say that choropleth er is strict but not mean it's strict because the names do have to match exactly but choropleth er tries to help you get your data into the right format and it does it with this data object called state dot regions if you look at this data set you'll see that it has a region column which is named exactly how regions are named in the map but it also has other columns for other common naming conventions for States the abb column is the postal abbreviation for each state and then there are FIPS codes that are in both numeric and character form if you're not familiar with FIPS codes i recommend that you google the term and read the Wikipedia article on it and here's a summary of what we've learned so far choropleth er works principally based on geography and the only geography that we've seen so far is States core plethora gives you example data in this case DF pop state the function for creating the map is state choropleth that function is in the package choropleth the map that choropleth er uses is state map and it's located in the choropleth ER maps package finally the metadata for the regions is called state dot regions we're now in a position where we can learn about all the maps the choropleth or ships with this is the full table that lists all of the maps one way to approach this talk would simply be to go through each element of this grid but if we do that there won't be enough time to cover the data integrations which i think is very important so what I'll do now is point out that there are basically two types of maps u.s. maps and international maps and I'm going to talk about one u.s. map the county map and one international map the countries map and I'll go through each element of the grid for those two geographies and then I'll just briefly talk about the other maps all of the US maps and choropleth are come from the US Census Bureau we're going to be principally talking about state and county maps and what I want to point out is that there's a strong relationship between states and counties which is each county is part of one and only one state zip code tabulation areas by contrast actually cross both state and county lines that never happens with counties here's the grid for County geography the example data is DF Pop County the function for creating the choropleth is County Korolev the map it uses is County dot map and the region metadata is county dot regions to learn more about how this works just look at the example data DF Pop County of course it has a region and a value column the region's though might surprise you there are no two US states that have the same name so we can use state names when making state corabeth's but there are 31 counties in America that are named Washington so names are not a unique identifier for counties instead we use FIPS codes and I won't be talking in detail about them but I'll say one thing for County FIPS codes the first one or two digits refers to the state the remaining three digits refer to the county within the state if you look at County dot regions you can see that of course there's a reagent column then the FIPS code appears as a character which basically means there's a leading zero and then we see the county name state name and the FIPS code for the state as well as the abbreviation for the state so again the data that you map will probably not be in the numeric FIPS code but you can always use the county dot regions data set in order to help chorister data into the format that choropleth requires if you look at the county map you'll see that there's also a region column and this is what your data will wind up being joined to in order to make the final map if you make a County choropleth and set num colors equal to 1/4 plethora we'll use a continuous scale and this makes it very easy to see outliers in the data in this case the dark blue county is Los Angeles Los Angeles is both the city and a county and it has over 10 million residents no other county in America has nearly as many people the county choropleth function also has a new parameter state zoom each function and choropleth ur lets you zoom based on the geography that you have for example the county correction lets you zoom by giving it a vector of counties but you can also zoom based on state and choropleth or will simply show all the counties within the state or states that you give it here we're using two colors which shows us which counties are above and below the medium here's the grid for all us maps that choropleth our ships with one thing I want to point out is that the zip code map is very large and so can't be stored on cran so it's in a separate package choropleth or zip which is available on github I'll also point out the zip codes are very complex map zip codes are maintained by the US Postal Service and they do not release a map this means that there is no official zip code map and so when I refer to zip code Maps I'm actually using zip code tabulation areas which are the Census Bureau's version of zip codes also census tract maps are very complicated the Census Bureau does not release a single national map of census tracts all census tract maps are by state and that's why there is no example data for them choropleth ER has two types of international maps administrative level zero maps at administrative level one Maps administrative level zero simply means countries of the world administrative level 1 refers to the first order administrative subdivisions of a country we've already looked at two administrative maps states and counties administrative maps just mean maps that show where legal borders are that show the limits of taxation and where laws can be enforced the reason why we use the term administrative level one is that different countries have different names for their administrative regions for example in America we use the word States Canada uses provinces and Japan uses the word Prefecture z' all the International maps in choropleth er come from this website natural earth data and if you want to learn more about these maps I encourage you to visit this website here's the grid for the country maps the example data is called DF pop country the function to make the map is country choropleth the map is named country dot map and the map is located in choropleth er maps and the region metadata is called country dot regions if we look at DF pop country we'll see what the example data looks like I want to point out again that regions are the most complex part of choropleth er an example of a problem with naming us states is Washington DC which is technically not a state but some people will call it Washington Comedy C where as the census map calls it simply District of Columbia the country map is an order of magnitude more complex because while here I write United Arab Emirates the official language of the United Arab Emirates is actually Arabic so they don't use English at all to name the country themselves if you look at the country dot regions data set you'll see that in addition to the region there's a code called ISO to see and you can google that term to learn more about it but the reason that I use that data set here is because choropleth er works with the World Bank's wdi data set and they use ISO to see for their naming convention we can create a map of this data by typing country choropleth with DF pop country when I look at this map I don't see any significant trends in the data although I will point out that the data set ranges from two hundred and forty seven thousand to 1.3 billion if we want to see where these outliers are we can set num colors equal to 1 this force is a continuous scale which makes it very easy to see the outliers and here you can see that both China and India have over a billion people we can see the map by looking at country dot map and I'll point out at the last column is the region column this is the full filled out grid one thing I want to point out is that the admin one functionality to create administrative level one Maps of countries works slightly different than all the other Maps first of all it's in a separate package choropleth or admin one second of all the add the first argument to the admin one choropleth function is the country that you on a map and the example data is just for one country Japan if you'd like to learn more about choropleth ER and
especially the mapmaking aspect of it and how the maps work and where they come from I recommend taking my course map making in R with choropleth er for example I mentioned earlier that zip code maps are very complex this course actually has three separate lessons on zip code maps and goes through all the details additionally each lesson has downloadable code which is very important because it allows you to explore an experiment in a structured way the most common question I get about choropleth ER is how someone can extend it to work with their own shapefile or their own map this often comes up when people are trying to map specific cities for example I live in San Francisco and San Francisco is divided up into what are known as supervisor districts in Sheep files for our programmers I go through an example project of downloading the shape file for the city of San Francisco by supervisor district I then download a specific data set noise complaints by supervisor district and I teach you how to tweak that map how to import it into our how to create a choropleth or object-- for it and finally how to make the final map now that we know all about the maps and court of leather we can look at the data integrations that choropleth or supports choropleth ER has three data integrations the first is for the Census Bureau's American Community Survey the second is for the World Bank world development indicators and the third is for the Bureau of Labor Statistics unemployment data this map is different than any other map we've seen so far first of all it shows per-capita income and not population second of all instead of showing a value for a particular point in time it shows the percent change in income between 2010 and 2015 I created this map after listening to a podcast with the author of a book that talked about the history of fracking in America he made a point of saying how North Dakota used to be a fairly poor state however after fracking was invented people realized that North Dakota sat on a lot of oil that could now be extracted and there was a tremendous boom in North Dakota I wanted to use choropleth er to verify the author's claim and I did and this map shows that he was correct between 2010 and 2015 the per capita income in North Dakota increased over 25 percent which is much larger than any other state I gave a talk at the American Community Survey data users group conference where I showed this map and also used choropleth er to create other maps to explore that phenomenon I now teach people how to create this map in my free course learn to map census data in R which you can sign up for at census mapping course com first I want to point out this new data set d/f state demographics if you look at it of course you'll see a column named region you'll also see columns called total population percent white percent black percent asian percent hispanic per capita income which we'll be looking at today median rent and median age this data all comes from the 2013 American Community Survey note though that there is no value column and this is because I expect the user to create their own value column to map the data that interests them to do this type the following code DF state demographics dollar sign value equals DF state demographics dollar sign per capita income now that you have a column called value you can create this map by typing state choropleth DF state demographics and you'll see that this data set ranges from 20,000 to 45,000 choropleth er also has a function called get state demographics if you give it a year it will give you the exact same variables that we've been looking at so far but for a different year in order to use this function you'll need to get a census API key which you can get for free from the Census Bureau here I'm getting data from 2010 by typing get state demographics 2010 and then I'm showing the first few values of that dataset note that this dataset has the exact same columns as DF state demographics but the values of course have changed we can calculate percent change like this DF 2010 equals get state demographics 2010 DF 2015 equals get state demographics of 2015 and then we need to set a value column on both datasets we can then calculate the percent change by typing DF change equals calculate percent change 2010 2015 this is a function that I wrote that assumes both data frames have a column called region it then joins those two datasets by the region column and calculates the percent change between the value columns of the two datasets and of course it returns a data frame with one column named region and one column name value we can then map that dataset by typing state choropleth DF change setting num colors equal to zero num colors equal to zero forces choropleth ER to use what's known as a divergence scale which is very useful for looking at percent change the value 0 which means no change is white large positive numbers are dark blue and large negative values are dark red here it looks like there are no large negative values although Nevada did have a small decrease in per capita income now I'd like to tell you how to learn more about census data the first point I want to make is that I have greatly simplified the underlying data for you if you look at get state demographics and simply type it with no parentheses you'll see all the code that is involved in getting that data from the Census Bureau one point I want to mention right now is that even though the columns called percent Hispanic it's actually percent Hispanic all races the way that the US Census Bureau analyzes race and ethnicity is actually quite complicated there's a rule I believe that any time the federal government asks someone for their race or ethnicity have to offer a minimum number of options here we can see within one race they ask if you are white black or african-american American Indian or Alaskan native Asian Hawaiian native or Pacific Islander or some other race and then there are options if you are two or more races even more complicated is that the Census Bureau does not consider Hispanic as a race they consider it to be an ethnicity this means that you can be Hispanic and any one of the options that I've already listed and this simply means that there's a great deal of complexity in the census data for race and ethnicity and when I say percent Hispanic it's actually percent Hispanic or Latino for all races you can learn more about how the Census Bureau treats race and ethnicity at this Wikipedia page if you'd like to learn more details about working with the American Community Survey in R and with choropleth er I recommend taking this free course that I created for the Census Bureau you can find the course simply by typing this title into Google mapping Census Bureau data in R with choropleth ur in module four which I call data details I go into much more detail about the American Community Survey when the our consortium learned about my course on census Academy they became interested in promoting our as a language for working with US census data they funded Logan Powell and I creating this document a guide to working with census data in our the Census Bureau actually has over 100 surveys and programs this document details the five most popular data sets and then gives you a link to see all of the data sets that they publish it also details the seven most popular are packages for working with US census data and then it links to all of the packages on cran that work with census data if you're interested in learning more about using our to analyze census data I highly recommend you read this document here's a grid that details how choropleth ur helps you work with US census data for each geography there is population data which you've already discussed each geography also has a similar object the details more demographic data we've already seen DF state demographics but there's also DF county demographics and DF zip demographics you can get the same data for other years by typing get state demographics get County demographics get zip demographics and get track demographics and finally although I don't have time to talk about it here if you want to get eight specific tables data you can use the function get ACS data and that will get you data for any geography that you're interested in now that we've talked about how choropleth er can help you get ACS data let's talk about how choropleth ER works with World Bank data the World Bank publishes a dataset called World Development indicators or wdi and they describe it this way the World Development indicators is the primary World Bank collection of development indicators compiled from officially recognized international sources it prevents the most current and accurate global development data available and includes national regional and global estimates luckily there's already an R package for working with this data set it's written by Vincent Errol bundok who is a political science professor at the University of Montreal if you'd like to learn more about this package I highly recommend visiting the packages github page all the data in the wdi dataset is identified by a code and I don't expect you to know the code of any data set off the top of your head but if you visit the wedi webpage they do provide search functionality which can make it easy to get codes also the our package wdi provides a search functionality for you to get codes based on keywords as well here I'm looking at the life expectancy data set and there's a new function choropleth er wdi it requires you to give it a code and a year and then it creates a map for you and here you can see the life expectancy at birth code at 2017 ranges from 52 years to 84 years and there are interesting patterns in the data set for example it appears that Central Africa is a cluster that contains a lot of countries that have a low life expectancy see also Western Europe appears to be a cluster of countries that have a high life expectancy if we set the number of colors equal to two we can see what countries are above and below the median value and here we can see that almost all of the countries in North America have above the median life expectancy similarly almost all the countries in Africa have below the median life expectancy and we also see that almost all the countries in Central Asia from India to Russia have below the median life expectancy finally if we set the number of colors equal to one the scale changes to a continuous scale this makes it very easy to see outliers in the data and it appears that two countries in Central Africa are outliers in terms of having a low life expectancy previously we've only seen outliers on the positive end for example Los Angeles County is an outlier in America for having a large population and India and China were outliers in terms of having a large population but here it appears that there are countries that are an outlier in having a very low life expectancy finally choropleth ER makes it easy to analyze and visualize data from the Bureau of Labor Statistics here's a map that shows the unemployment rate in US counties for 1990 and you can see that the data set has a very large range from 0.4 percent to 40 point 80 percent there are also some visual patterns to the data for example it appears that most counties in the central part of the country had a very low unemployment rate similarly most of the counties in Alaska seem to have a very high unemployment rate choropleth er gets this data from the Bureau of Labor Statistics originally I'd wanted to use their API but it had query limits that made it difficult to use so I scraped the data and packaged it the state data comes from 2000 to 2013 and the County data comes from 1990 to 2013 this data is in a separate package our unemployment data unfortunately shortly after creating the package the Bureau of Labor Statistics changed the format of their pages and so my script can no longer update the data the main data set that contains county unemployment data is DF county unemployment the first column is the region column then there's the county name state name and the year of data this boxplot I think shows why it is so convenient to have this data in our with a single line of code our will create a box plot for each year's worth of data what really stands out to me are two things the first is the number of outliers in the data especially how high some of the numbers get second you can really see the global recession that began in 2008 in fact there's a massive jump in the median unemployment between 2008 and 2009 we can create the map that we saw at the beginning of this section as follows the function is county unemployment choropleth it requires a year nineteen ninety and I set the number of colors to nine nine is actually the largest number of discrete colors that kora supports similarly there's a function state unemployment choropleth they will create state level kora plus of unemployment data finally I want to wrap up by telling you how you can learn more about choropleth er the main website for the project is choropleth er comm that URL actually redirects to re Lambton comm / open source I mentioned this because a number of users especially those who work for the federal government have difficulty visiting choropleth or comm I think that their security settings are somehow set to not allow you to visit a website that immediately redirects to another URL if you visit choropleth ER comm you'll start by seeing a brief description of the project also there's a sign-up form for my course learn to map census data in R as I mentioned earlier that course will send you one email every day for a week containing instructions to make this map which shows how per capita income changed in the u.s. choropleth er has also led to number of interest in consulting projects if you'd like to learn more about those projects and also to get help on your own project then I recommend clicking through to learn more I also provide instructions on how to get your questions answered I recommend asking all technical questions on the website Stack Overflow I especially recommend tagging your question with the word choropleth and if you click here you'll see examples of past questions I also run a Facebook group called our Maps and open data which is a community for a broader discussion about this topic if you have any questions or simply want to leave a comment I recommend joining and asking there on this page you can also learn more about the various courses I've created for choropleth err as I mentioned earlier map making an R with choropleth err is intended as the definitive guide to the project and it provides a lot of detail about how choropleth ax works and all of the maps that I have shapefiles for our programmers is very useful if you want to extend choropleth err to work with your own map if you purchase the are mapping bundle you'll get both courses at a discount and finally even though choropleth err is conceptually one body of work it's actually five different are packages if you click through to the documentation you'll get installation instructions for each package as well as each function and data object in each package this concludes the talk and now I'd like to just reiterate the main points in the introduction you learned the basics of using choropleth ER and how it supports exploratory data analysis then we went behind the scenes and I showed you what happens when you type the function state choropleth once you learned that I explained how that pattern applies to all the maps that choropleth or ships with finally I showed you the three data integrations that choropleth er supports and those are for the Census Bureau the World Bank and the Bureau of Labor Statistics and finally I taught you about choropleth or comm which is the best resource for learning more about choropleth er thank you
