A Finance Map of NYC:  Reducing Carbon and Driving Large Scale Energy Efficiency with a Public Database To Support PACE Lending

Fuzzy Matching is a technique used in computer-assisted translation as a special case of record linkage. It works with matches that may be less than 100% perfect when finding correspondences between segments of a text and entries in a database of previous translations.The concept of Fuzzy Matching is to calculate similarity between any two given strings. And this is achieved by making use of the Levenshtein Distance between the two strings. FuzzyWuzzy is a library of Python which is used for string matching. Fuzzy string matching is the process of finding strings that match a given pattern. FuzzyWuzzy has many powerful functions that allow us to deal with more complex situations such as substring matching .

We also tried to achieve Fuzzy Matching by Ruby. We used a geocoder gem to make a more educated guess by comparing the latitude and longitude of two addresses and returning true if they are close enough. It connects to more than 40 API worldwide so it is very powerful. We also used the address matcher library in Ruby to achieve our functions. Specifically, we first created an address library with all of the known address strings we wanted to match against, then called match to find the closest matching address. For instance, we can use our library to measure whether “1000 5th Ave, New York, NY 10028” and “1000 5th Avenue, New York City, NY 10028” is the same address 

We compared the owners with the largest number of buildings in New York City mentioned in the ’Who owns all of New York’ article .  "No. buildings / vacant parts" is the data mentioned in the article, and "owner in our merged file" is our own data. We found that our own data is quite different from those mentioned in the article. After analysis, we think that we mainly use the real owner data in DOB data, while the data used in the article is only  the  business  name  or  company  name , so we have more data.  Moreover, our data  do not count the buildings of the whole NYC, so these two reasons lead to great differences.

We need to organize different databases and constantly increase information about owners. So based on the data of the Pluto website, according to the information of the BBL column and the registration information of the HPD data column, we used Python to combine the data of HPD and Acris website. Without changing the order of basic data, we added new information about owners, including gas emission information, geographic information, tax information, mailing information and housing use information.After we found the DOB database, we added the real owner information. Our database already has real owner information, but now the problem is that a BBL will correspond to many different owners. We are summarizing data through different owner types and business names.So then we use SQL to make classification and adjustment, and classify the same BBL according to the house type.

We compared the owners with the largest number of buildings in New York City mentioned in the ’Who owns all of New York’ article . "No. buildings / vacant parts" is the data mentioned in the article, and "owner in our merged file" is our own data. We found that our own data is quite different from those mentioned in the article. After analysis, we think that we mainly use the real owner data in DOB data, while the data used in the article is only the business name or company name, so we have more data.  Moreover, our data  do not  count the buildings of the whole NYC, so these two reasons lead to great differences.