So many Americans! So few Britishers!

West Coast = best coast.

The median year of birth was 1983. I am glad I am slightly younger than average, even if the million people born in 1985 make me feel old! Also the relatively narrow range of births here is interesting, but I guess that’s just how social groups form.

So many males!

About 80% of respondents were straight, which is less than the incidence in the overall population, as far as I know? It’s a big homofest up in these comments.

Everyone be white though.

Just one person identifying with a disability.

This is probably the biggest difference compared to the overall population. About 88% of people could be classified as non-religious/non-spiritual, and of those, the majority are atheist, as opposed to more wishy-washy definitions. If there is a bias in things people share, it is almost certainly in this dimension.

…or maybe this one. The two “other”s did not sound like conservative beliefs, meaning there is no representation from the right side of politics here. Bunch of communists I swear

I’m not totally sure if the outlier to the left isn’t a typo/conversion error. In any case, most ladies appear to be right around the average height for the female American population, while males are slightly taller on average.

It looks like there’s not a lot of demand for Google Reader mixers.

I’m not sure if this is much of a difference to the rest of the population or not – can you expect about a third of people in the general population to not want children? In any case, there is plenty of representation from the people who love babbies, so keep sharing those birthing videos.

Almost everyone is employed, and of that group, almost everyone is employed full-time. Plenty of time to spend on Google Reader, I guess. I wonder in what proportion of those workplaces is Google+ blocked? :( Most of the people not in the labour force were full-time students.

Perhaps it’s not surprising that a social network revolving around sharing and discussing (mostly) long-form content skews towards the educated, but I thought it was interesting to see just how educated the group is. We’ll be going into more detail regarding how much that education is costing everyone as soon as I find some more time.

Type

As expected, full-fat milk is the most popular, making up 43% of all milk consumed over the sample period. Skim milk made up a further 34% of milk consumption. Soy (13%) and reduced-fat milk (7.5%) made up almost all of the remainder. Ryan used half and half a couple of times because he is gross.

Meals

Perhaps not too surprisingly, by far the most popular uses of milk were as a snack, or with breakfast. These two meals accounted for 79% of all milk usage. In contrast, comparatively little milk was consumed for lunch, dessert or non-meal uses like baking. This was very different to my experiences, where these plus snacks account for almost all of my milk usage.

Date

[Click image to enlarge]

The total volume of milk consumed was the highest in early January (at around 2.5 litres [85 oz] per day), but dropped off slightly to around 2 litres [68 oz] per day towards the middle of January as everyone flaked out. The amount consumed seems relatively consistent from day-to-day, while the amount purchased has a very lumpy distribution. This reflects the consistency of both peoples’ milk usage, but also their regular shopping habits.

Taking a look at the total amount purchased and consumed per weekday, it becomes obvious that most people shop for milk on a Sunday and/or a Wednesday. In addition, consumption is relatively consistent during the week, with a noticeable drop-off of approximately 15-20% less consumption on the weekend.

Time

[Click image to enlarge]

Consistent with the data on mealtimes accompanying the milk, the most common time to consume milk is from 8am-9am. This seems a little late to me, shouldn’t you lot be at work? Milk consumption then decreases with each successive hour to minimal usage from 12pm-3pm. However, from then on, it is clearly afternoon snack time. There doesn’t seem to be any uniform time that people take their afternoon snack, dinner, or dessert, but there is a spike at 10pm-11pm, suggesting that people like some milk in their tummies before bed.

The available data on milk purchases (which is a minority – most people didn’t bother noting what time they bought milk) indicates that most people buy their milk early in the morning or the middle of the day. This is likely to be buying things on the way from work or on lunch breaks, unless you are all old people that love to shop first thing in the morning on the weekend.

Milk Purchased and Consumed

The average quantity of container purchased was 2.2 litres [74 oz], and the average quantity consumed at once was 293 millilitres [10 oz]. The most popular quantity of milk to purchase was the half-gallon [1.9 l], which was purchased on 41% of occasions. Following this was the gallon [3.8 l], purchased on 32% of occasions, and the quarter-gallon [0.9 l], purchased on 10% of occasions. You guys sure do like buying huge cartons of milk!

The quantity of milk consumed in one sitting was obviously far more variable. While the distribution was wider, it seems to be bimodal, with the most popular amount to have at once being around 200 ml [7 oz] (mostly due to Jo consuming the same drink two to three times per day), and the second-most popular quantity being around 350 ml [12 oz] (Wes and Ryan’s preferred quantity). Quantities ranged from Gretchen splashing 30 ml [1 oz] of milk into a cup of tea, to Peter and Matt W drowning their troubles in almost a litre of milk (34 oz)! How do you even get glasses big enough to fit that much?

Milk Purchased and Consumed, by Person

[Click image to enlarge]

[Click image to enlarge]

Here is where the flaws of the data are most noticeable. On several people’s returns there are bunches of dates that are close together with a lot of entries (usually towards the start of the month), and then large amounts of dates with no entries at all. There are two possible interpretations to this: 1) people tended to start recording data diligently, and then gave up; or 2) people just don’t use that much milk. Given that this is the internet, I would put a lot of my money on the first explanation. However, the second is almost certainly also true to some extent, and we can’t really say to what extent either is true for individual people.

To resolve this, I decided to calculate milk usage in two ways. The first was the total milk usage you reported in the sheet, divided by 31 days in January. (Some people went above and beyond and recorded milk usage for more than 31 days; for these people, I calculated their milk usage using the first and last dates they entered.) This is your average daily milk usage, assuming you fully recorded every instance of a purchase or consumption of milk. It is represented by the blue bar. The second alternative measure of milk consumption is the average daily milk usage for only the range reported in your spreadsheets. For example, if someone consumed 1,200 ml from 17-19 January, and had no other entries, I took their consumption to be 400 ml per day. This measure applies if everyone stopped filling in their spreadsheet at some point, but continued with exactly the same pattern of consumption. It is represented by the blue bar added to the red bar. It is clear that each person’s true usage of milk in January/February should be somewhere in the range of the red bar.

It is at this point that we can say something about the likely relative flakiness of individuals. Peter, Ryan, Kristen, Wes, and Gretchen all have very small or non-existent red bars, indicating that they kept up the recording exercise throughout the entire period. In contrast, Casey, Davin, and Matt W scored relatively high on this proxy measure of flakiness. You might want to consider what sort of family you married into here, Gretchen.

One of the major reasons for doing this exercise was to see exactly how much milk we as a group tend to consume. From the collected data, we can see that there is a large variation between individuals, with some (such as Peter, Ryan, and possibly Jo and Matt W) consuming upwards of 500 ml [17 oz] per day, while others (Mike, Shara, Amy Berg, Laura, and maybe Casey) would struggle to finish that amount in an entire month. Overall, we consume (at most) about 200 ml [7 oz] of milk per day, which translates to about 73 litres [19 gal] per person per year. Using figures available here, we consume far less than the typical Scandinavian, Britisher, Australian, or Canadian. We consume slightly (about 10-20%) less than the average New Zealander or American. Interestingly, if we all drank as much milk as Peter, we would only be consuming slightly (about 10%) more than the average Finnish person. That’s a lot of milk!!

[Click image to enlarge]

We can also compare the reported purchases and consumption, and see how consistent they are with each other. To calculate this measure of consistency, I divided the average daily quantity of milk purchased by the average daily quantity of milk consumed. The resulting ratio gives an average number of litres of milk purchased for every one litre consumed. We can say that people with ratios well above one tend to be very wasteful (allowing your girlfriend to buy a lot of milk but not consuming it all), share their milk with others, or tend to underestimate the amount of milk they use. People with ratios well below one likely tend to overestimate the amount of milk they use, aren’t the household purchasers of milk (would it kill you to do the shopping once in a while Casey, Laura, and Matt W), or simply steal it off other people in the office (looking in Wes’s direction).

My Google Reader habits seem to be becoming more volatile (the right side of the top graph looks a lot more jumpy than the left side), which seems to have contributed to the average articles read for each day smoothing out a bit.

I steadily pick up new RSS feeds, and don’t tend to drop old or crappy ones as much. A lot of feeds stop updating, but I keep them around just in case.

As well as logging aggregate information, Google Reader also allows you to look at stats for particular feeds. However, it only gives you information for the top 40 feeds every 30 days. Because of this, the data isn’t very reliable: if a given RSS feed is the feed with the 41st highest traffic in all of my subscriptions, I have to pick it up as having zero posts. Still, it gives some indication.

Here are the biggest sites I’ve subscribed to:

The feeds I most frequently read the articles of:

The feeds I have shared the most:

Candy Consumed by Americans, 2000-2007

The average American consumes 24.5 pounds (11.1 kg) of candy (excluding chewing gum) every year.

Here is another what kind of candy Americans prefer. Americans categorise both chocolate and lollies into the candy family. It seems like they prefer chocolate.

Distribution of Candy Consumption, 2007

I also thought this was interesting:

Quantity of Sugar and Corn Syrup Used in Production of Candy, 2002-2006

In the period from 2002 to 2006, the amount of corn syrup used in the production of candy decreased by 3.8%. The usage of cane sugar decreased by 6.3% over the same period. Unfortunately they don’t publish this series anymore, so I can’t get more up to date information.

Summary Stats

Total spending: $13,821
Mean spending per day: $132 [-26%]
Median spending per day: $32 [-12%]
Most spent in one day: $2,475

One new graph: the distribution of spending by day.

The inner ring is the number of days I spend that amount of money, and the outer ring is how much money in total I spend on each of those days. So most days I don’t spend much at all (57% of days I spend less than $50, which would leave me with a reasonable amount of savings. But these days make up only 7% of my total spend. The problem really is at the far end of the distribution; the days where the total spent is more than $200. The average spend of these ten days is $908, and 10% of my days account for 66% of my total spending. So I tend not to spend that much, but when I do spend, I make very large purchases.

ps sorry if this updated like five times

Total Articles Read by Date

The slope in the middle of the graph is where i forgot to collect data for about a week.

Average Articles Read by Day

I still have no idea why Wednesday is my busiest Google Reader day. Possibly because I go to uni in the morning and then work late, and at about 6-7pm I zone out and start reading the internet.

Blue: Processing
Red: Making the pizza
Yellow: Cooking the pizza

This is awesome.

Update: The pizza was crap though.

Summary Stats

Total spending: $10,193
Mean spending per day: $179
Median spending per day: $36
Most spent in one day: $2,475

I haven’t been on holidays yet, so the holiday spending only includes the airfare so far. It’ll be interesting to see whether I spend more money in the long run on travel or on ridiculous electronics I don’t really need.

I categorised each of these expenditures into non-avoidable (rent, car, phone, food etc) and avoidable (clothes, electronics, holidays), to see whether I was truly wasting my money.

I feel this is skewed quite a lot, as I bought television ($1,800) and air fare ($1,600) within a relatively short space of time. In the period I have been recording, I saved only about 10% of my income. Usually I save around 30-50%. It does show that I am definitely spending too much money on crap I don’t need, though.

A definite pattern here. It’ll be interesting to see if this holds in the long-run (I suspect the Saturday-Sunday-Monday pattern will always be like this), or if it is skewed by the three days I made very large purchases (see the first graph).

A side benefit to keeping track of where I spend my money is that I also keep track of where I eat my lunch (I rarely take my own lunch to work because I dislike cold or reheated food and also I am incredibly lazy).

I’ve trying to be good, Subway and Thai food are relatively healthy right? I will probably be eating Indian more often because I found this awesome place that gives you three curries, a naan, poppadom, rice, and salad for about ten bucks. Ten bucks!