Data set use for analysis

In order to determine changes in home value differ between 1990 and 2000, a 1990 to 2000 (ten years) data sets were used.

## [1] 72693

Filter Urban Districts

## 
## rural urban 
## 12971 59722

Identify Common Variables

To enhance proper analysis of the data sets, common variables need to be identified for 1990 and 2000 data sets:

Table of descriptive statistics

The descriptive statistic below shows the summary statistics of all the variables used for analysis.

Statistic Min Pctl(25) Median Mean Pctl(75) Max
mhmval90 0 58,800 86,500 112,399 141,800 500,001
mhmval00 0 81,600 119,900 144,738 173,894 1,000,001
hinc90 4,999 24,355 32,033 34,160 41,520 150,001
hu90 0 916 1,337 1,382 1,804 11,003
own90 0 422 744 788 1,093 8,180
rent90 0 168 358 476 663 8,653
empclf90 0 1,015 1,555 1,625 2,157 11,816
clf90 0 1,101 1,664 1,732 2,291 12,497
unemp90 0 49 88 107 143 1,165
prof90 0 192 359 451 616 6,290
dpov90 0 2,219 3,250 3,332 4,365 23,619
npov90 0 112 248 406 526 5,890
ag25up90 0 1,444 2,121 2,187 2,865 18,461
hs90 0 626 1,064 1,141 1,559 12,383
col90 0 162 335 483 652 8,575
pop90.x 0 2,283 3,325 3,419 4,464 35,721
nhwht90 0 1,262 2,424 2,513 3,575 21,561
nhblk90 0 18 76 426 331 12,121
hisp90 0 27 75 347 280 13,873
asian90 0 9 31 111 94 7,899
p.white 0 64 87 74 95 100
p.black 0 1 3 12 10 100
p.hisp 0 1 3 10 9 100
p.asian 0 0 1 3 3 94
p.hs 0 69 74 74 80 100
p.col 0 10 18 22 30 100
p.prof 0 17 25 27 34 100
p.unemp 0 4 5 7 8 64
pov.rate 0 4 8 12 16 100

Exploration of Median Home Value

Initial conditions for data in 1990 and the home value for inflation.

Statistic Min Pctl(25) Median Mean Pctl(75) Max
MedianHomeValue1990 0 77,616 114,180 148,367 187,176 660,001
MedianHomeValue2000 0 81,600 119,900 144,738 173,894 1,000,001
Change.90.to.00 -660,001 -23,339 453 -3,629 19,100 1,000,001

Histogram of MHV

This histogram shows the distribution of home value changes between 1990 and 2000:

The distributions below compares the distribution of the data raw data between 1990 and 2000:

Change in MHV 1990-2000

The small initial values in the data are skewing the percentages. Assuming and average home value below /$10K is a vacant lot, these data can be removed. Outliers with over 500% change should also be removed to limit data skew.

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max.     NA's 
## -1.00000 -0.15942  0.00436  0.05019  0.20222 28.32257      145

How many entries show a >500% increase:

## [1] 26

Preview of the data with large increase (>500%) to see if this makes sense.

Plot the percent change variable:

With the outliars removed to minimize skew, we can now accurately represent the percent change in home value between 1990 and 2000.

Group Growth Rates By Metro Area

We often want to disagregate descriptive by some grouping in the data, such as metro areas. dplyr makes this easy by grouping then summarizing the data.

cbsaname ave.change.d growth
Corvallis, OR $73,403 75.42
Portland-Vancouver-Beaverton, OR-WA $67,900 71.8
Salt Lake City, UT $61,005 69.02
Boulder, CO $94,414 68.56
Provo-Orem, UT $60,131 67.84
Salem, OR $48,726 61.27
Eugene-Springfield, OR $51,072 61.11
Fort Collins-Loveland, CO $57,232 56.31
Longview, WA $40,726 55.06
Missoula, MT $45,520 54.35
Jackson, MI $28,210 53.41
Greeley, CO $41,416 51.69
Wenatchee, WA $45,676 50.45
Detroit-Livonia-Dearborn, MI $28,733 49.37
Yakima, WA $30,828 49.34
Ogden-Clearfield, UT $42,194 48.51
Denver-Aurora, CO $52,010 47.94
Monroe, MI $42,004 47.38
Bay City, MI $26,654 47.04
Logan, UT-ID $35,543 46.43
Eau Claire, WI $27,792 43.02
Madison, WI $42,897 42.14
Sioux Falls, SD $31,025 41.11
Grand Junction, CO $34,334 40.9
Mount Vernon-Anacortes, WA a $41,944 40.61

Comparing the median home value between 1990-2000 and 2000-2010. The above analysis shows that 1990-2000 has a median change of $453. Reviewing the tutorial analysis, 2000-2010 has a median change of $36,268. Change in home values have increased by over $35,000 in just ten years!

The average percent change in growth in home value between 1990 and 2000 is only 5% where the average percent change in growth in home value between 2000 and 2010 is 33%!

PART TWO

Measuring Gentrification

The original merged dataset we saved as d.full, so we do not need to reload it:

Recall our data steps thus far:

Min. 1st Qu. Median Mean 3rd Qu. Max. NA’s
-100 -15.94 0.4362 5.019 20.22 2832 145

Selection of Gentrification Variables

The selection of variables for operationalizing and defining gentrification: We added some variables from 2000:

Descriptive Statistics of Change Variables

The table below describes the summary statistics for the selected change variables.

Statistic Min Pctl(25) Median Mean Pctl(75) Max
mhv.90 11,712 77,880 114,312 148,728 187,440 660,001
mhv.00 0 81,600 119,900 144,738 173,894 1,000,001
mhv.change -660,001 -23,448 396 -4,122 18,968 963,701
pct.change -100 -16 0 5 20 2,832
p.white.90 0 64 87 74 95 100
p.black.90 0 1 3 12 10 100
p.hisp.90 0 1 3 10 9 100
p.asian.90 0 0 1 3 3 94
p.hs.edu.90 0 69 74 74 80 100
p.col.edu.90 0 10 18 22 30 100
p.prof.90 0 17 25 27 34 100
p.unemp.90 0 4 5 7 8 64
pov.rate.90 0 4 8 12 16 100
p.white.00 0 47 78 67 91 100
p.black.00 0 1 4 14 14 100
p.hisp.00 0 2 4 13 15 100
p.asian.00 0 1 2 5 5 95
p.hs.edu.00 0 67 72 72 77 100
p.col.edu.00 0 12 21 26 36 100
p.prof.00 0 23 31 34 43 100
p.unemp.00 0 3 5 6 8 100
pov.rate.00 0 4 9 12 17 100
metro.mhv.pct.90 1 20 41 45 68 100
metro.mhv.pct.00 1 20 41 45 68 100
metro.median.pay.90 14,871 28,906 32,457 32,924 35,833 52,374
metro.median.pay.00 23,012 39,457 43,139 45,054 49,522 73,701
metro.mhv.pct.change -99 -5 0 0 6 99
pay.change 4,930 9,775 11,441 12,130 14,001 26,211
race.change -100 -12 -5 -8 -2 100
metro.race.rank.90 1 20 41 45 68 100

Operationalizing Gentrification

## [1] 371
## [1] 17560
## [1] 0.02112756

By the above definition, based on the analysis only 2.1 percent of urban tracts experience gentrification between 1990 and 2000 as compared to 5.7 percent from 2000 to 2010.

17560 of census tracts are candidates, and 371 transitioned into advanced stage of gentrification.

PART THREE

The dorling cartogram shows the census data and groups from clustering.

## [1] 627

## [1] "character"
## [1] 627

The cartogram above shows candidates for gentrification in 1990 disaggregated by mean home value. The map shows a higher median home value in the North-West region of the city and lower median home values toward the center.

From 2000:

The cartogram above shows candidates for gentrification in 2000 disaggregated by mean home value. The map shows a higher median home value in the North and South-West region of the city and lower median home values toward the center.

The cartogram above shows the distribution of gentrification across the city between 1990 and 2000. The map shows that change in the median home value happens more toward the north and north-west areas of Pittsburgh.

The cartogram above shows that the percent change in median home value as a result of the distribution of gentrification happens in the far northern area of Pittsburgh.