SOURCE

This data was obtained from the Maricopa County Assessor under the search "Fast Food". The query has approximately 1342 results, with only 1000 returned due MCA Data Policies.

DATA CLEANING

Due to some Subdivision Name values posessing unescaped commas that interfered with Pandas' ability to properly align the columns, some manual cleaning in Libre Office was performed by me.

Aside from a handful of Null values, the data is fairly clean and requires little from Pandas.

NULL VALUES

Here are the sums and percentage of NULLS in the dataframe. Interestingly, there are 17 NULLS that do not have any physical addresses. This amounts to 1.7% of values for the Address, City, and Zip, and are all corresponding rows for those missing values.

I have looked into a couple of these on the Maricopa County Assessor's GIS Portal, and they do not appear to have any assigned physical addresses. This is a good avenue of exploration for EDA. Possibly an error that could be corrected, or some obscure legal reason, but interesting nonetheless.

Additionally, there are 391 NULLS in Subdivision Name accounting for 39.1%. This is a feature that I am interested in exploring to determine if there are any predominant groups. It could also generate a list of Entities that can be searched later to see if the dataset can be enriched beyond it's initial 1,000 record limit.

There are 348 NULLS in the MCR column. This is the definition according to the MCA Glossary

MCR (MARICOPA COUNTY RECORDER NUMBER)
Often associated with recorded plat maps.

This seems to be an uninteresting nominal value, so I will drop this columns.

While Property Type and Rental have no NULLS, 100% of those values are Fast Food Restaurant and N (for No), and therefore offer no useful information, and will be dropped.

I will leave the S/T/R column, although it also seems to be uninteresting nominal values, I am curious if there are predominent groups, and since it also has no NULLS, might be useful for further data enrichment.

The 1995 Land Use coverage was created as a joint effort of MAG (Maricopa Association of Governments) and MAG member agency staff. Land Use components were classified into 24 categories. The 1995 Land Use coverage is used for a variety of planning purposes including socioeconomic forecasting and air quality modeling.

An extensive survey was conducted to assess the species diversity of natural desert vegetation around the CAP-LTER study area. Remnant patch habitats (mostly parks and preserves) occurring within the urban matrix / fringe were emphasized. Three outlying areas (White Tank mountains west of town, Union Hills north of town, and Usery Mountain Park east of town) were surveyed for comparison. The original intent was to provide data on the woody and spring / summer herbaceous species in each area, but drought conditions allowed the spring herbs to be surveyed twice (1998, 2001) and summer herbs surveyed only once (1999). Thus, all locations have woody data but only a subset has herbaceous data. Surveys consisted of transects (elaborated below) summarized by the patch or outlying area name followed by a number (e.g. Unhill-11, Squaw-40, Hayden-8). Herb datasets are signified by a ‘P’ for spring and a ‘U’ for summer following the transect number (e.g. PaPk-1P, PaPk-1U).

This study examines the effects of land use on microclimate along several commercial-to-rural land use transects in the greater Phoenix metropolitan area.

Distribution of Praseodymium concentration in lichen tissue collected as part of a study of heavy metals in lichens in Maricopa County, AZ.