Lending Club

LendingClub is a US peer-to-peer lending company, headquartered in San Francisco, California.[3] It was the first peer-to-peer lender to register its offerings as securities with the Securities and Exchange Commission (SEC), and to offer loan trading on a secondary market. LendingClub is the world’s largest peer-to-peer lending platform.[4] The company claims that $15.98 billion in loans had been originated through its platform up to December 31, 2015.[5]

LendingClub enables borrowers to create unsecured personal loans between 1,000 and 40,000. The standard loan period is three years. Investors can search and browse the loan listings on LendingClub website and select loans that they want to invest in based on the information supplied about the borrower, amount of loan, loan grade, and loan purpose. Investors make money from interest. LendingClub makes money by charging borrowers an origination fee and investors a service fee.

LendingClub also makes traditional direct to consumer loans, including automobile refinance transactions, through WebBank, an FDIC-insured, state-chartered industrial bank that is headquartered in Salt Lake City Utah. The loans are not funded by investors but are assigned to other financial institutions.

The company raised $1 billion in what became the largest technology IPO of 2014 in the United States. Though viewed as a pioneer in the fintech industry and one of the largest such firms, LendingClub experienced problems in early 2016, with difficulties in attracting investors, a scandal over some of the firm’s loans and concerns by the board over CEO Renaud Laplanche’s disclosures leading to a large drop in its share price and Laplanche’s resignation.(source: Wikipaedia)

The Data

The Data comes from Lending club: https://kenyandataonline.herokuapp.com/dataset/1/14/. For those with more powerful computers, the full dataset is in Kaggle:https://www.kaggle.com/wendykan/lending-club-loan-data/downloads/lending-club-loan-data.zip/1 but fair warning the data has 700MB, so its a heavy dataset. The data has variables on the people that got loans from the company as well as the status of the loan at the end of the period: whether it was paid or not. The variables include details such as income,home ownership etc as well as loan details such as type of loan, purpose of the loan, loan interest and such like details. The full dataset: These files contain complete loan data for all loans issued through the 2007-2015, including the current loan status (Current, Late, Fully Paid, etc.) and latest payment information. The file containing loan data through the “present” contains complete loan data for all loans issued through the previous completed calendar quarter. Additional features include credit scores, number of finance inquiries, address including zip codes, and state, and collections among others. The file is a matrix of about 890 thousand observations and 75 variables. A data dictionary is provided in a separate file.

The idea is to analyse the data via explorarory analysis(descriptive statistics ) to get a better understanding of the data for furthe analysis or more advanced modelling that is beyond the scope of this analysis.

EDA

Numerical Variables

For the Numerical variables, a histogram will show you the distribution at a glance. of course the best start should be the summary statistics that show the mean,median,mode,max and min as well as other statistical measures of central tendency as well as of dispersion. Here we just skip this step and show some visualizations since people fing graphics more intuitive and compelling than tables of data and figures. Most data has the bell curve distribution which is distribution of most data close to the mean and the median with ow standard deviation as shown by the first three plots, the last one has more variation but this is expected with a variable such as income. What is interesting is that you will see that the company policy is to fund loans that are way less than the average income, and considering that the loans are mostly for around 3 years, the loans are affordable.

I found the interest rates to be an important data variable which is shown below. for this data we have a boxplot that shows the distribution better with the mid line being the median iterest rates. Here we find that most people are charged 12% on loans with some outliers in the data which amy be people with low credit score who would be considered more risky and therefore charged more interest. 5% is the lowest interest charged on any loan which is quite cheap for any loan especially here in Kenya. The median installments paid on the loans was 275 The box siply shows that fifty percennt of the population pay loans in istallments that are between 175 and 425.

##    int_rate
## 1:    10.65
## 2:    15.27
## 3:    15.96
## 4:    13.49
## 5:    12.69
## 6:     7.90

Non Numeric variables

The other variables are either characters or factors which are almost similar in their evaluation. Here we will plot them using piec charts so that we can see how the data looks. One can still use bar charts or even stack plots to show more detailed information but for now the piec charts will suffice. Although for the advanced statistical analysis, we will need to convert the variabes into numeric datasets using dummy variables ( or One-hot-encoding in Machine learning). But once the data is converted it is less intuitive to intepret onced graphed hence the visuals are made form the raw data.

relationship between variables and Loan status(eg correlation test $ plot it)

A correlation analysis looks at the movement of the variables relative to each other. what we are interested is the movement of the numerical variables in relation to the loan status. The correlation heatmap shows positive correlation though not very strong for all the variables.

cross tabulate the laon status vs loan purpose,loan status and grade

we can cross tabulate the loan purpose and the loan statust. here we want to discover if the reason for taking the loan influences the ability to pay.What you are more interested in are the fullypaid and charged off columns.Debt consolidation seems to be the main reason most people take loans but there is also a significant part of this group that is unable to pay their loan. If we used the full loan data, we may find a better pattern than the subset we are using but the results still look pretty insightful.

Another interesting analysis would be how credit scores and loan status relate. A cross tabuation of the two variables reveals that Grade A and Grade B has the most fully paid loans which may mean that the credit score positively influences ability to pay the loans or it may also mean that the company offers more people with grade A and B loans and therefore its logical they also pay more.The results are showed below on a heat map that helps spot interesting data patterns faster than just presenting data.

Top 40 employment title

It would be interesting to know from which companies and job titles do most of the company clients come from. we do a simple frequency analysis for this data. It is important to understand however that this type of information falls under natural language and therefore a mode detailed analysis (NLP) that tokenizes the data into words and then removes stopwords and further aggregates the common words would offer better results. However for this analysis we wish to remain simple just to give an idea of how to analyse such information.

what this means to the reader

The exploratory analysis of the data has given us a wide range of information on the type of users of the data as well as the company policies on loans and its MO.Armed with this knowledge, we can now create more complex analysis on the data to answer more advanced questions.