Normalizing Data in ArcGIS Desktop

okay let's explore some additional functions one of which is quite simple to use and quite powerful and that is the whole idea of normalizing your data let's talk about and demonstrate how to do that we've seen in other videos how easy it is to classify data and how easy it is to make different kinds of maps ranging from single symbol maps as we've got here to graduate the color maps to graduate a symbol Maps pie chart Maps density maps and more let's talk about something that is fundamental to spatial analysis and to statistics in general and that is normalizing data in this case we've got some state populations here so what we can do here is we can make a quick map showing the population of the states in 2010 there it is okay well what if I want to know the change in population from let's say 2010 from the ten years before from the year 2000 well that's easy to do because we've got this normalization field what what that does is that divides whatever value you have in the in the value field by the value that's in the normalization field so if we divide population 2010 by population mm hmm let's see what that gives us if we think about it if a state had let's say ten million people in 2010 and only five million people in 2000 we'd be dividing ten by five so two would be our result so a positive number over one would indicate that the state is growing in population a number less than one means the state is declining in population and looks like we do have at least one state that's declining in population let's go ahead and do that okay so this is the growth it indicates the growth doesn't it because these states are most rapidly increasing in population out west and down south in Texas as well as some south east and these states up here are either growing very slowly or else they're decreasing in population I'm really interested in this that decreased in population or States so I think I'll go over here and set my range to one for the breakpoint that means that whatever stayed here that have got symbolized in yellow will indicate to me that the state decreased in population right because we're dividing 2010 by 2000 okay let's go ahead and find out which one that was ah okay it looks like it was Michigan so if we hit the identify button go up here to Michigan and pull up the attributes for Michigan let's go ahead and investigate the data in 2000 the population was nine point nine three eight million in 2010 it was nine point eight eight million so it looks like Michigan is the only one according to this data set from census data 2000 and 2010 that decreased in population why is that what were the economic variables that are impacting the state what were the other variables that impact the state over that period of time why did it decrease if we went back to the 1940s and 50s maybe the height of the in migration due to automobiles we might be able to see a whole different story in fact let's do that it's pretty easy to do and we'll be able to use the same procedures we just explored so let's go ahead and divide the 1960 population by the 1950 population to try to get a handle on in migration to Michigan okay so 1960 divided by 1950 looks like there's even some states that decrease during the 1950s because I do have some states less than 1 and similar to what I was doing before let's go ahead and set the breakpoint to one and that way I've got some states that are decreasing I could even change the label here decreasing in 1950s okay and the rest are increasing and let's make that symbol yellow just so it will stand out all right so it looks like during the 1950s a couple of south central states Arkansas and Mississippi how do I know that can't I label these hey Arkansas and Mississippi decreased in population while Owen looks like also West Virginia while other states increased in population so it confirms our hypothesis let's go ahead and turn those labels off that we had a radically different situation back in the 1950s while look at Florida quite a bit of increase in the 1950s the West as well as we see today but there were some things going on in the 50s that were different from today how could we dig into the historical and geographic component of that change well GIS allows us to open the door to inquiry now remember with this normalization tool you just need to use some caution make sure you're dividing apples and apples for example if you've got some census data for example let's say you've got females in the value field and then you've got let's say mobile homes in the normalisation field in other words you've got population in the value field that in your normalization field you've got housing units specifically mobile homes if you divide females by mobile homes the computer you know being a computer it'll do whatever you specify but you're not your well in this case you're dividing apples by oranges so the result is not females and mobile homes see what I mean so you'd need to use caution and really understand what your data is is telling you and also where it was gathered from and who gathered it and why it was gathered so know your data and use caution when you're dividing things that in any type of data analysis you need to you really need to understand your data okay but that being said the normalization function is really powerful and wonderful what about population 1850 / 1840 now why are we going in Reverse why can't we divide 1840 by 1850 well we could we get different numbers just as long as you understand what those numbers are indicating you're fine in this case if we divide the the later year by the prior year by the prior census year that is we're going to see a negative number if a state decreased if we go the opposite progressed see a negative number if it increased so it's a little bit easier in my mind to go this way all right it looks like in the 1840s we didn't have any state that decreased in population we have some states that increased a huge amount nine times so let's go ahead and make those how about some sort of a hatch pattern really makes them stand out okay so these are the fastest growing states in the 1840s and okay why don't we have anything out here well there weren't there those states weren't existing in eighteen in the 1840s and so ah this looks like Wisconsin increased the most in the 1840s makes sense Wisconsin became a state in 1848 and you could look back in time and look at when states actually became states that probably coincided in many cases with their most rapid growth so let's go ahead and change this to the 1830s now so I'm going to go to 1840 / 1830 and let's go to a five classification method here and let's once again make this fastest-growing one some sort of hatch pattern and okay another variable that we've got in here is something that I classified and created and that is the change year by year so if we can if we map change any particular year we don't want to normalize it because we don't want to divide a rate of change by a total population so this is a case where normalization is not suitable because I already have a change component inside this variable okay so here's the change during the 1930s interesting we've got some states that that obviously it decreased quite a bit here 7% is what this is we've got another state on the high end that increased 36% let's go ahead and and change that and you know what I think I want to make my I want to make my states that that decreased I'm going to once again do what I did before I'm going to make a zero here in this case now we've got states that decreased it's going to be let's make that a bright yellow and then let's make this one a sort of a pale yellow okay so we've got yellows for the decreasing states in the 30s and we've got Reds for the increasing okay well I've got the dustbowl right there don't I I also have an interesting situation over here in Vermont what went on in Vermont during the 1930s let's go ahead and hit the identify button and just dig into the data a bit more so 1930 population of Vermont 359,000 and look 1940 359,000 looks like it only lost 400 people okay so it decreased slightly however we go over here to these dates take a look at Kansas in 1930 it had 1.88 million in 1940 it had 1.80 million so looking at these decimal places here going from looks like it lost a good 80,000 people in the 1930s what about Oklahoma scroll down there 1930 it had two point three nine six million in 1942 point three three six million so oh it looks like 60,000 people left Oklahoma during that those years fascinating okay so we've done some simple but powerful things here we have normalized and symbolized and started asking some questions about the data some of which we could answer immediately others we will have to do a little bit more research but that is the essence of geographic inquiry asking questions gathering your data investigating your data analyzing your data and then displaying your data and then acting on that and asking new questions Thanks

Channel: ESRIEdTeam

Views: 12,649

Rating: 4.9272728 out of 5

Keywords: Normalizing, Data, ArcGIS, desktop, ArcMap, 10, Esri, GIS, ratios, analysis, software

Id: JXuOJsoDa0c

Length: 12min 2sec (722 seconds)

Published: Thu Jul 28 2011