Author Archives: viorica

565 days of my life in numbers

If you’re like me, you’re often wondering: “What do statisticians do in their non-professional lives?”. Since this was one of those rare occasions where Google didn’t know the answer, I’ve decided to find it on my own and after three years here’s the answer:

Getting to this answer was tedious work. On December 8th, 2010 I’ve started logging every step I made, day and night, rounding the intervals to the nearest 5 minutes. My calendar records looked something like this:

My calendar

All this time I lived in Galway, Ireland. I kept logging until August 31st, 2012 with a break between July 17th and August 20th, 2011 when I went to Moldova and most of the time I was visiting family and friends. On July 30th my son Adam was born and for a month I tried to keep logging my daily routine. But it was impossible as practically I didn’t have a routine, all my activities were interrupted and more important I was so tired that I wasn’t able to keep accurate track of what I was doing. I decided to stop and didn’t include this last month data in the study.

New baby in the family brought us lot of joy, but also deprived me the possibility of working without being interrupted, so I postponed the analysis of my calendar for more than a year. Meanwhile I moved with my family to San Francisco, California. After we adjusted to the change and my baby became more independent I decided to start working on my project, kind of a New Year’s Eve 2014 resolution.

After taking away the visit to Moldova and the last month of logging I was left with 565 days. Applying a bit of Python magic, I exported the data from iCalendar into a Google spreadsheet. I corrected the misspelings and combined some activities until there were left 25 categories.

If I were to spend each of the 565 days the same then the activities of my daily living would be distributed according the following chart:

If I take away the sleep time wich accounts for 32.3% of my time, then my rest of the day has the following distribution:

These numbers were calculated from summing up all the activities I did during this period and then deviding by 565 days. Of course my days weren’t all the same, some activities I did perform daily, some weekly, some monthly and some occasionally. More details follow below, along with some charts to visualize the data.


“Evening” routine includes 48 minutes for bathing my kid, story time, tucking in bed and 29 min for my shower and bedtime preparation.

“Morning” routine was divided into school mornings and no school mornings. On weekdays we would spend 70 minutes from getting up until going to school and 67 minutes during weekends. The result surprised me, I was expecting more time doing nothing in the morning in weekends, but I guess the weekend lazy mornings in bed are a rarity when you have kids.

I had lunch and dinner at home 5 to 6 times a week and my average lunch average duration was 26 min, dinner 35 min and tea/coffee/snack breaks 19.5 min. I was busy cooking about 10 times per week with average duration 34 minutes.

Every day I had at least 3 breaks of 21 min, usually time in between/ interruptions from my main activities or activities not important enough to record.

16 times per week, 29 min long, I did cleaning and tidying up the house, laundry, cleaning after parties or kids’ activities.

12 times a week I would spend 33 min in front of my computer, mostly managing my mail, online enquiries and appointments, research, reading etc.

In this period I was mostly SAHM but I did some work on a freelance basis. I was busy working 4 to 5 times a week for 95 min long. My work-related activities consisted of:
– Tutoring college students in calculus, algebra, statistics and programming. In total 52 sessions 134 min long on average.
– Organising sewing and crafts workshops for kids during holidays. In total 25 workshops during this period with average duration 187 min.
– Volunteering for Sea2Sky science exhibition: 1h for training and 9h 45 min during the day of the exhibition.
– The rest is working for Sea2Sky and Galway Science&Technology festivals as statistician, statistical consultant on freelancer.com, preparing the tuition sessions and the workshops and writing on my blog.

During this time I took some courses with Open University and Udacity.com, so 4 to 5 times weekly I studied, 77 min on average.

I had 5 times per week of entertainment (movies, shows, Youtube etc.), on average 68 min long.

4 to 5 times weekly 72 min were dedicated to Eva: nearly 2 hours for karate (including the time to and from), nearly 3 hours for swimming (usually we spent some time after swimming in the coffee shop over there). Once during Christmas holidays we spend 12.4 hours making a 1000 pieces puzzle in 9 sessions 83 min long. Other regular activities were homework, playdates, board games, arts and crafts, birthday parties. One important mention: here isn’t included the time spent watching something together, going in town/library/church and other outdoor activities.

4 to 5 times a week I did “Other” activities for 71 min on average:
– About 45 hours of my time was Christmas related starting with writing cards to the family and friends and ending with throwing out the tree.
– I went to the doctor/dentist about once every two months either for me or my child, and about twice per month during my pregnancy. One visit lasted on average 70 min long and included the time taken to and from the doctor.
– I spend 121 hours preparing for my driving test, driving on my own, driving lessons and the two actual tests (yes I failed the first one).
– During this time we did a very important project in our live, we bought our first house. I was the one staying at home and doing “nothing”, so I was the one mostly involved in all the activities around it. I spent 289 hours (12 days) starting with enquiring about mortgages, scheduling viewings, hiring the solicitor and contactors and ending with unpacking and arranging our new home.
– I kept a log on all the food I ate on caloriecount.com for a month at the beginning of my pregnancy to see whether I get all the nutrients I needed. I spent on that on average 26 min daily.
– 24.6 hours of my life were wasted “counting sheep” at night. Over these 565 days I recorded 23 cases of insomnia, average length 64 min. I compensated that with 27 sessions of napping/resting, 42 min long.
There were some more “other” activities, but not important or interesting enough to be mentioned.

Breaking down the “Socialize&Communication” category resulted in:
– about 2 times per week phone and Skype conversations (mostly family and friends), 29 min on average.
– 2 to 3 times per week socializing with friends in my house or theirs, 66 min on average.
– about 2 times per week I spent in the company of my husband and sometimes family and friends staying overnight, 46 min on average.
– about 2 times per month conversations on serious topics with my husband (like our kids’ future), 58 min on average.

I did grocery shopping about 3 times per week, 52 min on average.

I went to town about twice a week, spending there 116 min on average for just walks, shopping, to a restaurant or coffee shop and attending special family events.

I recorded 165 sexual encounters over this period which results in 2.04 times per week, 13.3 min long.

About once a week I went for walks 97 min long. About once per month I went to the beach spending there 174 min, but considering only the months April to September results in twice per month. I did 13 sessions of physical exercises at home 33 min long, 11 sessions of yoga 80 min long (including the time to and from the venue), once I spend 2h playing tennis and once 4.8 h playing golf.

At least once per week I was sewing, sometimes as hobby and sometimes as work, 77 min on average.

About once a week we had family time for 81 min. Usually the three of us were playing a board game or doing something fun like going to the swimming pool. One mention: here isn’t included the time going to the park/playground/beach/cinema, watching a movie together or going in town.

 

I spent 18 min on average on the way to school and 37 min from the school. I have been in the school for some special events and volunteering (Parent Teacher Association), about 3 times a month during schoolterm speding 104 min on average.

I read 2 to 3 times a month for 38 min long. These were the rare moments when I was reading at home curled on the sofa and don’t include reading in the library, bus, waiting rooms or during karate and swimming lessons.

Breaking down the “outing” data results in:
– 1 to 2 times a month going out in town for drinks (164 min) or cinema (181 min)
– 1 to 2 times a month going to the library 121 min on average
– about once a month going to the church 139 min on average.


During this time I spent 212.5 hours (8.85 days) away from home. Averaging we were going away about once per month for 482 min (8 hours). The rest is time spent on the road. I recorded only 9 times preparing for the one day trips, on average it took me 62 min for packing and preparing snacks.

28 min daily went into “unaccounted” category due to occasional small gaps in my data logging.

And finally 5 to 6 times per week I spent 13 minutes logging all these activities.

Now as a SAHM I have the answer to the question “So what did you do today?” But most important I have an idea how to manage my time better, the most important resourse. And from a point of view of a statistician, this analysis is just a quick insight of what I can do with my data. iCalendar contains enough details to do a more thorough statistical analysis containing not only averages but distributions and correlations of my daily activities, timeline series.

The numbers show that I had a pretty good work-life balance in Ireland. San Francisco is a very interesting city, but I have this constant feeling of busyness and exhaustion, everything takes lots of time and energy: shopping, doing errands, my kids’ activities, school. I decided to start another project and now I am collecting data on how I spend my days here, so that I could compare my daily living in the two cities. After a couple of months I will be able to have some preliminary statistics. Meanwhile please share my article if you liked it and feel free to leave comments and suggestions about what else would be interesting.

 

Sea2Sky

A great science event, family oriented, takes place all over Europe across more than 300 cities since 2005, on the fourth Friday of September. This is Researchers’ Night and it’s about showing what researchers really do for society, in interactive and engaging ways.

Ireland participated twice, in 2011 and 2012. Both times Galway city hosted the event and NUIG were the organizers. The chosen subjects were sea, sky and space and the event was called Sea2Sky.

Sea2Sky attracted up to 10 000 visitors and the organizers were interested to assess the promotion and the impact of the science event and to rate the public interest in these subjects: sea, sky, space. The assessment was done through surveys, which I distributed and collected on the day of the event. In 2011 about 500 surveys were distributed and 378 collected and in 2012 about 600 surveys were distributed and 248 collected. The surveys were similar, except a few questions.  The incentive for answering the questions was the chance to participate in a draw for an iPod in 2011 and for a gift voucher in 2012. It looks like the iPod worked better as an incentive, with a response rate of 75%. The surveys can be downloaded below:

Sea2Sky survey 2011 (681)
Sea2Sky survey 2012 (935)

 

A bit of statistics theory before getting to the reports:
The analysis is based on significance testing: a number called significance probability or p value is calculated which quantifies the extent to which the data cast doubt on the null hypothesis –  the lower p value, the more evidence that the results didn’t occur by chance. In the analysis I used the level of significance α=0.05 and the null hypothesis was that there is no difference in answers between genders or age groups. Thus significant difference means that the results didn’t occur by chance, if the p-value is less than 0.05. Note that the p-value is not the probability of the null hypothesis being true, in other words p=0.02 doesn’t mean that the probability of finding a difference occurring by chance is 0.02, but we have conditional probability:p=0.02 is the probability of obtaining a difference as the one calculated from the data, assuming there is no difference.

Further I will present a few interesting results. The full reports can be downloaded bellow:
Sea2Sky statistical report 2011 (719)
Sea2Sky statistical report 2012 (829)

In both years the proportion of females in the sample of respondents is greater than the proportion of males. In 2011 over 80% of the respondents and in 2012 nearly 60% were of school age . This is not surprising as before 2pm the main visitors were the booked in schools. There is no significant difference between genders within each age group.

First question was about where they found out about Sea2Sky the participants could choose more than one answer. Main source of information was school, as expected by the large number of school age respondents. Adults listed nearly equally school, newspaper, saw a sign and other sources (at work).

When asked to list the reasons for attending the event most of the respondents mentioned at least one reason. The distribution of answer is presented on the following diagrams:

In 2011 there was an interesting result. Only the answer “I am/want to be a scientist/engineer” produces significant difference between genders: 30.5% of males and 12.% of females are or want to be scientists/engineers.

The respondents were asked to express in their own words what they liked most or least about Sea2Sky festival and what else would like to see in the future.


In 2011 the most popular subject (20.2%) was Sea, double than the Space(10.1%) and in 2012 the situation reversed and Space (21.9%) was the most popular, Sea(12.6%) ranking third. In 2012 doubled the proportions of the respondents that mostly appreciated the various exhibitions and the interactive experiments. In both years about the same proportion of respondents appreciated the child-friendly atmosphere at the festival. For more details see the reports.


When asked what they liked least both years about the same proportion of respondents answered that they disliked nothing or like everything (41.1% in 2011 and 40.5% in 2012). The visitors mostly disliked the crowds, noise, heat, queues. Some didn’t like experimenting/feeling the seaweed. Some complained that there wasn’t enough time for so many activities in one day. For more details see the reports.


In 2012 most of the respondents reported that next year they would like to see even more science/experiments/shows. Others would like to know more about space and NASA discoveries, technology, robots, submarines, etc. Some would like just more entertainment and free or less expensive refreshments. More details to be found in the reports.

 

 

Females and males are about equally interested in Sea and Sky. The 2011 report indicates that males(71.%) are more “interested a lot” in Space than female(55.9%). The 2012 report indicates that children are more “interested a lot” in Sky(46.6%) and Space(72.6%) than their parents (37.2% and 43.9% correspondingly).

The organisers of the event were particularly interested how the visitors, especially the younger ones, interacted with the scientists/researchers, how they felt about participating and whether they got a better understanding of science. The following diagrams show the distribution of answers in 2012. The answers in 2011 were similar.

 

 

 

 

 

There are more questions about the interest in sea, sky, space and how well the organisers covered these subjects, also whether the respondents see improvement in 2012 comparing to 2011 in terms of location, parking, volunteers, cafes, crowds and queues. For more information download the reports following the links above.

Finally the respondents were asked whether they would come next year. About 98% said “yes ” in 2011 and about 95% in 2012. One curious thing is that just about 30% of the respondents in 2012 have been to the festival in 2011. It remains to see how is going to be 2013 festival.

Another science festival is taking place in Galway each November: Galway Science and Technology Festival founded to bridge the gap between young people and industry and encourage students to study STEM (Science, Technology, Engineering and Maths) subjects at third level. The event takes place over two-week period, during which a lot of science shows are brought to schools throughout Galway city and county. The festival is ended with a Final Day Exhibition at NUI Galway. In 2011 I prepared a survey similar to Sea2Sky which was distributed among the visitors of the exhibition. The report based on 516 surveys, can be downloaded, following the link below:

Galway Science&Technology statistical report 2011 (796)

Moldovenii vor unire cu Rusia?

Pe data de 1 martie 2011 Centrul de Investigaţii Sociologice şi Marketing „CBS – AXA” a publicat rezultatele unui studiu sociologic desfăşurat în cadrul „Monitorului Euro – Asiatic”. Sondajul investighează aspecte economice, politice și sociale din Republica Moldova. În particular prezintă interes răspunsul la întrebarea „Dacă aţi decide integrarea Republicii Moldova, cu care din următoarele ţări, după părerea dvs. ar fi mai corectă unirea ţării …”. Conform răspunsurilor respondenților s-a estimat că 40% din populația Moldovei dorește aderare la Rusia.

Rezultatul sondajului a fost preluat de diferite organizaţii mass-media și publicat online. Reacţia cititorilor a fost promptă și plină de indignare, mulți fiind de părerea că într-un fel sau altul sondajul a fost falsificat.

Quo vadis?

CBS – AXA menţionează că respondenţii au avut posibilitatea să aleagă una sau mai multe din opțiunile prezentate (răspuns multiplu), de exemplu Rusia şi România. În imaginea alăturată, fiecare bară indică frecvenţa alegerii ţării respective de către respondenţi. Normal, suma acestora poate să depășească  100% (de fapt este 126%). Ca rezultat nu putem trage concluzii despre frecvența sumară a mai multor opțiuni. De exemplu aceleași persoane care au ales Ucraina, Belorusia și alte țări din CSI, puteau sa aleagă Rusia in același timp. Deci e incorect de spus că 68% din respondenți vor sa adere la UE sau Rusia. De fapt, prin metode deductive se poate de estimat că frecvențele sunt distribuite astfel: „Rusia și alte țări din CSI” – 40%, „Uniunea Europeană, inclusiv România” – 28%, „Independență” – 28% și „Greu de răspuns” – 5%. Sumar 101% (cîțiva respondenți au ales mai mult de 2 opțiuni).

Analiză

Eşantionul a fost de 1030 de persoane, realizat în 64 de localităţi ale Republicii Moldova în perioada 28 ianuarie – 8 februarie prin selecţie aleatorie a gospodăriilor şi persoanelor intervievate. Pornind de la premiza că sondajul într-adevăr a avut loc și nu e bazat pe creativitatea unui număr mic de persoane, are sens de văzut ce ne mai spun datele, pe lângă graficul precedent.


Întrebarea de mai sus a mai fost adresată și indirect: „Dacă aţi avea posibilitatea de ales, in care din ţările sau grupurile de ţări, aţi fi dorit să trăiţi?”. Se poate de sumat frecvențele pentru Rusia, CSI și fosta URSS și obținem 36.9%. Am menționat mai înainte că vor unire cu Rusia și alte țări din CSI 40% din populație.

 

 

 

Comparație dintre răspunsurile la întrebarea despre “Unire” și “Trai”. Se observă o corelație strânsă între dorința de aderare la o țară și dorința de a trăi în țara respectivă.

 

 

 

 

Prezintă interes în particular graficul alăturat. Din el rezultă că moldovenii de etnie doresc unire cu Rusia în proporție de 30 la sută, rușii – 58 la sută, ucrainenii – 73 la sută, găgăuzii – 64 la sută și bulgarii – 80 la sută.

 

 

Ultimul grafic este util și pentru a verifica dacă CBS – AXA a efectuat corect stratificarea eșantionului pe etnie. Conform recensămîntului din 2004 (excluzând Transnistria) distribuția pe etnii este următoarea: moldoveni și români – 78.4%, ucraineni –  8.3 %, ruși – 5.9 %, găgăuzi – 4.4 %, bulgari – 1.9 %. Aceste proporții trebuiau respectate și la selecția participanților.

Deci ideal eșantionul trebuie să fie distribuit  în modul următor: 0.784 proporția de moldoveni, 0.059  cea de ruși, 0.083 cea de ucraineni, 0.044 cea de găgăuzi și 0.019 proporția de bulgari.  Am menționat mai sus care sunt proporțiile corespunzătoare fiecărei etnii în ceea ce privește dorința unirii cu Rusia: 0.3 din moldoveni,  0.58 din ruși, 0.73 din ucraineni, 0.64 din găgăuzi și 0.8 din bulgari. Ca rezultat sumar pentru întreg eșantionul obținem proporția de 0.3734 care corespunde la 37.34 % ceea ce este destul de aproape de 40% și demonstrează că CBS – AXA a luat în considerație corect etniile.

Concluzie

O remarcă despre rezultatele alegerilor din 28 noiembie 2010 la care PCRM a obținut 40.6% din voturile numărate la secțiile de votare de pe teritoriul republicii. În opinia mea rezultatul sondajului corelează cu rezultatul alegerilor în republică. În același timp se pare că CBS – AXA a omis un grup reprezentativ de cetățeni – moldovenii aflați după hotare (doar 6.75% au votat au votat PCRM).

Datele arată că CBS – AXA  a realizat corect sondajul în ceea ce privește etniile și supoziția că populația Moldovei a fost supra-reprezentată de minorități nu este adevărată. În cee ce privește ponderarea pe vîrstă și regiuni agenția nu ne oferă detalii. Referitor la unirea cu Rusia consider 40% din populația totală din republică o valoare plauzibilă. În același timp se pare că doar 30% din moldo-români își doresc această unire. Iar pentru cetățenii de peste hotare această cifră probabil este și mai mică.

 

Alcohol consumption: Moldova vs Others

On 11th February 2011 the World Health Organization (WHO) published the Global Status Report on Alcohol and Health 2011 and earlier on 5th January 2011 they published the European Status Report on Alcohol and Health 2010. Since then on a lot of websites circulated the idea that Moldova is worldwide No.1 drinking country with an average consumption of 18.2 litres of pure alcohol per person per year. After reading some comments, mostly expressing deeply the injustice of WHO towards Moldovan people, I decided to download both reports and have a closer look at the data and see what’s hidden behind this number.

How much is 18.2 litres of pure alcohol?

First of all I want to make clear what means 18.2 litres of pure alcohol in more common terms as beer, wine and spirits. According to WHO the following guideline is used for the conversion into litres of pure alcohol: 5% for beer, 12% for wine and 40% for spirits.
So 1 l of pure alcohol corresponds to 20 l beer or 8.33 l wine or 2.5 l spirits. Using multiplication we obtain the following amount of drinks per Moldovan person per year: 364 l beer or 152.7 l wine or 45.5 l spirits. Divide by 52 and we get the corresponding amount per week: 7 l beer or 2.92 l wine or 0.875 l spirits. In order to top up to this amount if you are Moldovan you have to drink weekly about 14 beers or 4 bottles of wine (a regular bottle of wine has 0.75 l) or 29 shots of spirits (a regular shot is 30 ml).
Health agencies of various governments have different maximum alcohol intake recommendations. In UK, for instance, the maximum recommended weekly amount for men is 210ml of pure alcohol, and 140ml for women. This is equivalent of about 10 beers for men or 7 beers for women. Thus, Moldovans are exceeding British safety level by roughly a third.
Ok, we admit it: we do drink a lot, but let’s see how is the situation in other countries, particularly neighbouring countries Romania, Ukraine and Russia and Europe as a whole.

Data interpretation

WHO uses several key indicators for describing each country’s drinking profile: types of beverages, consumption of recorded and unrecorded alcohol per capita (APC), average alcohol consumption among drinkers (not abstainers), abstainers (lifetime and former drinkers), alcohol use disorders and alcohol dependence, fatal road accidents, patterns of drinking.

So, what total alcohol consumption per capita (APC) means? APC is obtained by summing the entire amount of alcohol (recorded and unrecorded) consumed in a country during a year and divided by total adult population (15+ years).

But not every adult drinks, there is a proportion of persons that don’t use alcohol at all (lifetime abstainers)  and then there is the proportion of adults who used to drink in the past, but because of different reasons did not drink in the last year (former drinkers).  Here is a chart with the abstention rate.

You can clearly see that Moldova has the lowest rate of abstainers. We do drink, but all together. I will introduce an observation from Global Status Report on Alcohol and Health:

“Lifetime abstention is the opposite of per capita consumption. In countries with high APC, there are fewer lifetime abstainers compared to countries with low APC. This suggests that WHO regions with the highest per capita consumption commonly found in the European Region or other developed areas, do not necessarily have the highest consumption per drinker. Often, per drinker consumption is particularly high in countries with moderate or even low APC combined with high abstention rates In such countries, the relatively smaller percentage of the population that drinks is consuming alcohol at high levels.”

In order to make a better estimate of alcohol consumption you have to extract the population that doesn’t drink at all or didn’t drink for the last year from the total population and you get the total population who drinks.  And then you divide the total alcohol consumption over a year to the total population who drinks and you get the average consumption per drinker.

Below is the chart of alcohol consumption per drinking capita.

Yeah, Russians drink more than Moldovans, and so do Romanians and Ukrainians. Actually let’s not celebrate we aren’t far behind.

After reading the reports I collected the data about alcohol consumption among drinkers in 39 European countries. Only Bosnia and Herzegovina, Turkey, Ukraine, Portugal, Russia, Croatia, Andorra and Romania drink more that Moldova. The drinking score per capita among drinkers vary from as little as 5.48 litres pure alcohol in Israel to 55.99 litres in Bosnia and Herzegovina (this value is unusually high even comparing to second larger score Turkey with 29.42 and I think it require further investigation). Moldova is No 9 in this top with an average intake per year of 23.39 litres of pure alcohol per capita; it’s still a lot of alcohol, but it is of some consolation that we are lower in the top comparing to Russians and Ukrainians.

And finally let’s see what Moldovans, Romanians, Ukrainians, Russians and Europeans like to drink.

Moldovans nearly equally drink wine, beer and spirits. Europeans drink a little bit less wine, Romanian less beer, whereas Ukrainians and Russians don’t like the wine at all and drink much more beer (Ukraine) and spirits (Russia).

WHO data sources and methodology

Both reports use data provided by Global Information System on Alcohol and Health (GISAH).

The recorded consumption comes from production, import, export, and sales data often via taxation.The unrecorded consumption is based on European Survey on Alcohol and Health that was sent to all Member States of WHO in 2008.

I didn’t find any document entitled Global Survey on Alcohol and Health on WHO website, but I found the general WHO guideline to surveys: World Health Survey. And on ec.europa.eu – European Commission website I found a document regarding WHO’s survey about alcohol entitled Global Survey on Alcohol and Health and Global Information System on Alcohol and Health.

Unfortunately, there is no information published about how and who conducted the survey in Moldova. On Moldova’s profile on WHO website there is no information regarding this. I sent them an email inquiring about the survey and I am still waiting for an answer.

I prepared an analysis with full-description of the key indicators, along with a more detailed description of data sources, survey, possible errors and more tables and graphs.  For those interested it is available for download here:
Full analysis report (2544)

Conclusion

YES, Moldovans do drink a lot, more than most of Europeans, however less than Russians or Ukrainians and in a less harmful way. And I have to outline that this analysis is as reliable as the data provided by European Survey on Alcohol and Health.

So anyone up for a drink?