Spending audit, 2016


Mint tracks all of my spending. Mint automatically categorizes transactions into categories like “Pharmacy” or “Rental Car & Taxi”, and I periodically check and change Mint’s categorizations.

I exported the data from Mint and, using Pandas, assigned each transaction to categories that were less granular and more relevant than Mint’s categories, like “transportation” or “work expenses”.

Then I classified transactions into an even coarser set of categories, classifying transactions by their purpose, like “food,”, “entertainment”, or “charity.” But many transactions serve multiple purposes. For example, a $12 restaurant meal provides maybe $3 of food and $9 of enjoyment. An MBTA pass can get me to work and then to a party.

So I assigned some of the granular categories to multiple coarse categories. For example, I buy lunch at Harvard partially because I need to eat and partially so I can spend more time working instead of cooking, so I assigned ``work food” 50% to “food” and 50% to “work expenses.” I assigned restaurant spending 25% to “food” and 75% to “amusement.” However, it was impossible to assign a purpose to many transactions. Spending on transportation serves many purposes, so I just left it as “transportation”. Cash could have been used for anything, so I lumped it in with “unknown.”


I live with my husband in a one-bedroom apartment in Cambridge, MA. I’m a Harvard grad student. I earned money by teaching at Harvard in the spring, from Harvard for no reason in the summer, and from working part-time at a data science start up since the summer. [1]

I got married this year. The wedding cost less than $1000 [2], but people gave us money, and we actually came out ahead.

I bought an expensive laptop this year, which was by far the largest expense after food and rent. I also bought a lot of clothing, since I’m at a point in my life where I shouldn’t be wearing shirts from high school and ripped-up athletic shoes to work. I expect to spend less on clothing in the future. I spend a lot on convenience; since I’m paid hourly, it’s a better financial deal for me to buy an $8 lunch than to cook it, pack it, and clean dishes myself. I think I could save time and money by cooking a huge amount of food once a week and pre-packing lunches, but I can’t ever seem to make that work.

I made $46,572.50 this year. Since I’m filing jointly it’s hard to say what I paid in taxes, but it’s about $6,800, leaving an after-tax income of about $39,800.

My biggest expense after rent and taxes is usually charity, and I usually donate at the end of the year. I took the Giving What We Can Pledge several years ago: I’ve pledged to donate everything I make over $34,000 post-tax, and at least 10% of my income. ($34,000 in 2012 in 2012 dollars, $35,264 today.) That left me obligated to donate $3,757 in 2016. I donated my “2016 donation” of $14,400 to GiveWell in the beginning of 2017 rather than at the end of 2016 for tax reasons (and I’ll donate again at the end of the year, and call that my “2017 donation”). I’m not going to count that here, though.

Spending by category (granular)

Category Amount Number of transactions Average per transaction
Rent 10,350 12 862.50
Taxes 6,800    
groceries 2691.47 87 30.94
laptop 2350.38 1 2,350.38
clothing 763.23 12 63.60
restaurants 705.56 45 15.68
work food 672.88 78 8.63
utilities 492.40 21 23.45
work expenses 470.42 6 78.40
sports 353.06 6 58.84
cash 290.50 7 41.50
transportation 232.83 17 13.70
unknown 215.62 6 35.94
CVS 196.15 23 8.53
other necessities 188.76 2 94.38
books and magazines 170.57 19 8.98
software for education/work 167.73 4 41.93
hair 165.00 2 82.50
amusement 148.47 15 9.90
wisdom tooth removal 141.80 1 141.80
gift 133.88 7 19.13
appearance 84.00 4 21
home improvement 50.75 2 25.38
coffee 39.54 11 3.59
charity 35 1 35
website 13.75 2 6.88
other 1.47 4 0.37
wedding -503.79 15 -33.59

Spending by category (coarse)

Coarse Category Corresponding Granular Categories Examples
entertainment amusement, 75% of restaurants Google play music, Felipe’s
wedding wedding venue, food, alcohol, music
self-improving entertainment books, magazines, sports The Economist, rock climbing, trail races
appearance clothing, hair, appearance  
education/work/professional development laptop, website, 50% of work food, noise-canceling headphones, Stata  
other necessities coffee shops, other necessities, CVS soap, instant coffee, phone
charity charity none (in 2016)
unknown unknown, cash, other ATM withdrawal
food groceries, 50% of work food, 25% of restaurants food
medical wisdom tooth removal  
transportation transportation Uber, bicycle repair, train
housing rent, utitilities, home improvement  
gifts gifts birthday presents
taxes taxes W2 withholding
Coarse Category Spending
housing 10893.15
taxes 6,800
education/work 3338.72
food 3204.30
appearance 1012.23
entertainment 677.64
self-improving entertainment 523.63
unknown 507.59
other necessities 424.45
transportation 232.83
medical 141.80
gifts 133.88
charity 35
wedding -503.79

This is a total spending of $27,408, or about $20,608 if you don’t count taxes. I usually spend about $18,000 a year, not counting taxes, so this was an average year before the new laptop. I used Betterment to invest the rest, putting as much as possible in a Roth IRA.

[1] I develop econometric algorithms. It’s a lot of fun. Ask me about it.

[2] Pretty good for 38 people if you ask me!