Tag Archives: quantified self

Using RescueTime to Answer the Question: When Do I Write?

I‘ve written quite a bit about how much I write, and that I find ways to write every day. What I haven’t talked about much is when I do my writing. Remember, I have a full time day job, and two little kids, so my time is very limited. In order to write every day, I needed to adjust to the fact that I couldn’t count on a fixed time of day, or a fixed duration of time in which to write. I had to learn to write whenever the time became available.

Since early this year, I have been using RescueTime on all of my computers. For those not familiar with it, RescueTime is an application that tracks what you do on your computer and provides you with data about your productivity. RescueTime has a database of applications and websites and ranks them anywhere from Very Unproductive to Very Productive. It gives you a “productivity pulse” from 0-100 telling you how productive your days were. One of the things that RescueTime captures is when you used an application and for how long.

Yesterday, I finally got around to playing with RescueTime’s API, and was able to pull out data about my use of Google Docs, which is where I do all of my writing. I learned a lot about when I write by looking at that data. I also confirmed some things that I already knew.

When do I write each day?

I went back to March of this year and took data from March 1 through yesterday. I filtered the data to look at just the “Writing” activity in RescueTime–that is, applications that are related to writing. I further filtered those down to Google Docs to ensure that I was capturing my regular daily writing, all of which I do in Google Docs. I aggregated the data by hour to see when during the day I typically do my writing. Here is the results:

When Do I Write

You can see that the vast majority of my daily writing is done between 7-9 pm. Indeed, of the 102 hours of writing that RescueTime has logged since March, more than half–53.1 hours–has taken place between 7-9 pm.

This data aggregates all days in the date range, including weekends. I sometimes write early on Saturday or Sunday mornings, especially when I know I have a big day ahead, so there is a spike there. Also, I sometimes write when the kids nap on the weekends (something that is increasingly rare), and so you see a small spike between 1 – 3 pm as well.

If I break things down by weekday/weekend writing times, here’s how things look:

Weekday-End Writing

The pattern on weekends is roughly the same as weekdays, and yet the chart is a little deceiving because it makes it look like I write a lot less on weekends than on weekdays. But remember, there are only 2 days in a weekend, and 5 days during the workweek. From March to the present, I’ve spent 76 hours writing during the workweek. This averages out to 15.2 hours per weekday.

However, I’ve spent about 30 hours writing on weekends in that same period. This averages out to 15 hours per weekend day. Put another way, I’ve written a total of 15 hours for each day of the week in the period from March to the present. I write just as much on the weekends as I do on weekdays.

An important note on the data collection

One thing I want to emphasize here. All of the data was collected automatically by RescueTime. I did not have to log my writing time. I did not have to take any extra steps, beyond opening my document and typing. RescueTime captures it all, automatically. Indeed, there is a gap in the data. When I was on vacation in Maine, I took a laptop with me that I forgot to install RescueTime on, so none of my time that week was captured.

But the key point here is that all of this data was generated without my having to take a single action beyond doing what I normally do. Of course, I did have to spent a few minutes creating the charts for this post, but the data was already there. I just had to grab it and process it. This, in my opinion, is a vital element to consistent tracking. If I had to take steps to log my time each time I was writing; if I had to “clock-in” and “clock-out” I would never have collected the data in the first place.

Reminder: My Google Writing Scripts are Available on GitHub

After my inaugural post for The Daily Beast appeared, I’ve been asked almost daily if the scripts I mentioned in the post are available. They are available on GitHub. I put them there last July. I hesitated to mention them in the post on TDB because I didn’t want to come across as promoting my own stuff. But since I’ve been asked almost daily since the post appeared, I’m thinking that maybe I should have. Ah, well. If folks are interested in trying out the scripts, or improving upon them, you can access the code on GitHub. Be sure to read all of the instructions there to get them working correctly.

Writing Stats for the First Half of 2014

Hard to believe it is already July. Half the year is over and that means I have writing data for the first half of 2014. Here is what the first half of the year looked like:

Writing 2014 Jan - Jun

In case it isn’t clear, the blue bars represent each day’s worth of writing (word count). The yellow line is the goal I set for myself to try to reach each day. It is secondary to getting any actual writing done, but I find it useful to have. The red line is the most useful when looking at trends. It represents the 7-day moving average word count. You can see that the first 3 months or so were kind of spiky, moving up and down around the goal, but beginning in April, when I started in earnest on the second draft of the novel, things picked up. Indeed, since mid-April, I haven’t written less than 500 words on a given day, the longest span I’ve gone doing that. Further, you can see that the 7-day moving average has not only exceeded my daily goal, but has been more than twice that goal for the better part of the last three months.

I thought it would be interesting to compare the first six months of 2014 with the first six months of 2013. I didn’t start my daily writing regimen until late February 2013, so there is some data missing early on. But here are the two years compared:

Writing 2013 and 2014

The red is 2013 and the blue is this year. You can see that since about May 2014, I’ve been regularly outstripping that same period last year.

In the first 6 months of 2013 (including the 2 months before I started my daily writing) I wrote just under 100,000 words. In the first 6 months of 2014, I’ve written 163,000 words. While I have also tracked my writing time using Rescue Time, I haven’t parsed that data in any detail. That said, I have worked out a model for estimating time from words written. 163,000 words comes out to about 109 hours spent writing in the first 6 months of the year. That’s 4.5 days worth of writing. Not very much in the grand scheme of things, but then again, with the full time job and two little kids, my time is fairly limited. It averages to about 906 words per day, or roughly 35 minutes per day of writing.

I have only a few deliverables to show for all of this writing so far: a couple of nonfiction articles. Otherwise, the writing has been working toward two projects: the second draft of my novel, and a novella that I’ve been working on intermittently when I need a break from the novel. So my published words in 2014 is a tiny fraction of what I’ve written. But I expect that to change as the second half of the year unfolds.

New Writing Record Yesterday!

I wrote just over 1,000 words yesterday, which is par for the course these day. It was my 305th consecutive day of writing, and my 448th out of the last 450 days. Both are records that I keep expanding with each passing day, but yesterday, according to my Daily Almanac, I set a new record:

Goal Record

I have hit my pre-set word count goal for 34 consecutive days, beating the old record of 33 days. My system allows me to set a daily word count that I try to aim for. For me, it is 500 words. Keep in mind that this is a secondary goal for me. My real aim is to write every day. Sometimes, it is less than 500 words, but that’s okay, so long as I am writing every day. However, for the last 34 days now, I’ve written at least 500 words every day. Here is what the last 34 days looks like for me:

34 Day Record

During the last 34 days, the blue bars have exceeded the purple goal line every day. You can also see that my 7-day moving average (the red line) has been above 1,000 words/day for two thirds of that time period. I’m cooking with gas, as they used to say.

This 34-day period coincides with me starting on the second draft of my novel and in that sense, there is a little bit of an illusion here. I’ve struggled with getting the opening just right and have reworked a lot of material during this time, so that while I’ve written quite a bit every day, the novel draft is still in its beginnings, although things are definitely getting better. This is something that I don’t worry about in the first draft, where I am telling myself the story. But in the second draft, I need to make it something that will be interesting to a wider audience, and that is more difficult, especially the beginning.

During the last 34 days, I’ve written 36,758 words of fiction, which his pretty darn good for someone who only writes 30-40 minutes/day. It is not, however, the most I’ve written in a 34-day period. That record came on August 28, 2013 when, in the prior 34 days, I wrote 49,288 words.

Still, I’m pleased with my progress. While the novel is slow to get moving, one of my real goals is to be able to write more every day, and that means squeezing more time out of the day to write. I can write 500 words in 20 minutes. In the last 34 days, I’ve been averaging 1,081 words/day, which means I’ve been steadily expanding the time I have available to write. Most full time writers I know aim for 2,000 words of copy each day. I’m about halfway there and I think that is a good thing.

My First 30,000 Step Day

I‘ve had a FitBit device for more than 2 years now. I average between 15,000 – 20,000 steps/day. I’ve gotten my 25,000 step badge, but the 30,000 step badge has always eluded me. Not anymore.  Yesterday evening, as I turned back onto my street from a long evening walk, this happened:

30K Day

Kelly and the kids were out, and the street outside the house was desolate because it is being repaved this week, so I celebrated my achievement alone:

30K Celebration

As you can see, 30,000 steps is about 14 miles. I only have detailed records going back 2 years, but I think it is safe to say that this is the second best distance I’ve walked in a day in my entire life. I think the top day took place sometime in 1999 or 2000 when I walked what I estimated to be about 15 miles in Manhattan, wandering about for most of the day. Still, I’ll take the 30,000 steps.

By the time I went to bed last night, I had amassed 31,194 steps for the day, which is my new record, and will likely remain so for some time to come. It is really hard to get 30,000 steps packed into a day.

FitBit emailed me a note of congratulations, letting me know that I’d received my 30,000 step badge. But there was also a little hint of challenge in that note:

30K Badge

Really? Another 5,000 steps? I’m going to be happy with my 31,194 steps and leave it at that.

Remembering Everything: An Evernote Coda to My Life In Weeks

Yesterday, I wrote about looking at my life in weeks, based on this excellent post over at Wait, But Why. In looking at my life in weeks, I began to wonder about how much of it was documented. Here is the chart I produced yesterday, for reference. Each square represents one week of my life since birth.

My Life In Weeks
Click to enlarge

I decided to take a slightly different cut at this. I looked at when I started keeping a paper diary; when the paper diary was replaced by blogging; and when I started using Evernote to “remember everything.” Here is what the same chart looks like with those three areas of self-documentation highlighted.

My Documented Life in Weeks
Click to enlarge

As of this week, my life has been 2,197 weeks long. Looking at this data, a fair chunk of it is documented in one form or another. If you take all three methods listed above, it totals 950 weeks, or 43% of my life. What this means is that I can go back to any of those weeks and find, to a pretty good degree of accuracy, what I was doing. I have the documentation in one form or another.

The blog continues as well (obviously) and that has been going steady now for 444 weeks, or 20% of my life. The blog doesn’t contain as much detail as Evernote does, but it is still a better overall source of documentation than my paper diaries ever were.

However, the documentation doesn’t get really accurate until I started using Evernote and attempted to “remember everything,” in part by going paperless. I started using Evernote in December 2010, which means I’ve been using it for 178 weeks. That amounts to just 8% of my life. That said, that 8% of my life is well more documented by far than the previous 92%. I can go back to almost any day in the last 178 weeks and tell you virtually everything about the day, from the weather, to what purchases I made, to what words I wrote on a specific story, to what new things my kids were doing.

I think that this is pretty cool, and I think that it will make for something special for my kids to be able to look through when they are older and I am much older. The details of life are often lost in memory, but when you can capture those details as discretely as I’ve been able to do with Evernote, very little is lost and it paints a vivid picture of “what life what like when Dad was a (relative) youngster.”

Fun with FitBit Data: Seasonal Activity on Weekdays and Weekends

I thought it might be interesting to take a look at all of my steps in the last year or so, but breaking them down into seasons and weekends vs. weekdays. I’ve done just that in the charts below. These charts1 are not composites of my daily walking. They are total steps for the seasons on weekdays and weekends. Each bar represents a 5-minute interval, and when you see one such interval with 4,000 steps, that is across the entire season, not a single day. Still, provides some insight into daily patterns, and especially differences between those patterns on weekdays and weekends, as well as seasonal differences.

Summer 2013

Weekdays

Summer 2013 - Weekday

Weekends

Summer 2013 - Weekend

Fall 2013

Weekdays

Fall 2013 - Weekday

Weekends

Fall 2013 - Weekend

Winter 2014

Weekdays

Winter 2014 - Weekday

Weekends

Winter 2014, Weekends

I did not include the spring of 2014 as we are only partway through the spring and the without a complete set of data, there is no means for comparisons with other seasons.

A few observations:

  1. While summer and fall are relatively close on weekdays, there is a big difference between summer and winter.  On peak summer days, my morning walks totaled more than 4,500 steps in each 5-minute interval over the course of the summer. For that same time in winter, the number was barely 3,500 steps, a thousands steps less in each 5-minute interval. The weather plays a big factor in how much walking I do between summer and winter.
  2. The patterns in the weekday data is consistent, even though the numbers vary from season to season. I am creature of habit when it comes to my walking.
  3. Patterns are virtually nonexistent on weekends. About the only consistency I see is a low step count around 3 pm across all seasons. I suspect this is because we’re typically home at this time, and the kids are napping.

I’ll try to remember to post a follow-up when I have a complete data set for the spring. Although I suspect the patterns for weekdays will look much like the Summer and Fall.

 

  1. The data comes from some Google App scripts I have that pull by minute-by-minute steps data from FitBit using their API. The data was crunched and the charts were generated using Mathematic.

My Life in Weeks

Yesterday, I saw a fascinating post that posed an interesting way to look at the events in one lifetime. The post, entitled, “Your Life in Weeks,” provided an interesting graphic of what a lifetime looks like when broken out into weeks. It identified several phases of what would be typical for someone living in the United States. I was fascinated by this approach, and so I attempted my own version of it, with an emphasis on my writing life. Here is the result, and below are notes keyed to the numbers in the chart. Note that the 0 year, 0 week at the very top left is my birthday, so the weeks and years are relative to that, as opposed to the New Year.

My Life In Weeks
Click to enlarge

Notes:

  1. Started Junior High School.
  2. Started High School.
  3. Started College.
  4. Started writing and submitting stories (January 1993)
  5. Graduated from college (June 1994)
  6. Started at my day job (October 1994). The light purple shading represents the time I’ve spent at my day job, the same job which I hold today. And let me help you with the math: this October will be 20 years.
  7. My first story sale (January 2007). Note the gap between the time I started writing and submitting stories (#4) and this first sale.
  8. The Little Man is born.
  9. The Little Miss is born

All of the orange squares represent stories or articles that I sold and were published in various magazines, so you can see them in relation to that first sale as well as other events in my life.

The red square in Year 42, Week 13 is today.

I really like this way of looking at the events that take place in your life relative to one another. I used a Google Spreadsheet to create the weeks of my life and it was surprisingly easy to do to.

Enjoy these posts? – Tell a friend

Recommending readers is one of the highest compliments you can pay to a writer. If you enjoy what you read here, or you find the posts useful, tell a friend! Find me online here:

Twitter | Facebook | Google+ | Blog | RSS

Or use one of the share buttons below. Thanks for reading!

FitBit Experiment: Measuring Battery Life from Low-To-Empty

Yesterday, I performed a little experiment with my FitBit Flex. When I arrived at the office, I received an email alert that my FitBit Flex battery was low. My charger was at home, and I certainly didn’t want my Flex to miss counting any of my steps, but I decided that this was an opportunity for an interesting trial. I would see how many steps (and how many hours) the Flex would last before giving up the rest of its stored power and shutting down. I received my email “low battery” alert at 7:59 am and had just 1,275 steps so far that day.

FitBit Battery 1

I decided that I would not alter my routine at all, but go through my normal process, getting my daily walks and periodically checking to see how the battery was doing. And that is exactly what I did.

Just before 3 pm, 7 hours after the low battery notice, the FitBit app started showing my battery as “empty.” At this point, I had just about 9,000 steps total.

FitBit Battery 2

I resigned myself to eventually losing some steps, but figured it was worth the sacrifice in the name of science in order to find out just how long the battery lasted after the initial “low battery” email notification. So I continued with my day.

By the time I finally went to bed last night, my Flex had not yet died. The battery still showed up as “empty” despite the fact that I had put a total of 17,450 steps in.

FitBit Battery 3

That was enough for me. Rather than put it into “sleep” mode only to have the device quit on my sometime in the night, I decided to charge it overnight. I had enough data to answer the important question.

The results of my little experiment can be summed up as follows:

  • Time from low-battery message to “empty” battery indicator: <= 7 hours
  • Steps from low-battery message to “empty” battery indicator: ~7,200
  • Time from initial “empty” battery indicator to when I decided to charge: ~7 hours
  • Steps from initial “empty” battery indicator to when I decided to charge: ~8,000

And here is the answer to the question that I was really trying to get in the first place: If I receive a low battery message, how long can I use my Flex before it will lose power?

The answer, based on yesterday’s experiment, is

  • At least 14 hours
  • At least 15,500 steps

Seems to me that is useful information. Next time I get one of those alerts, I know I don’t have to rush back home to grab my charger. I can go virtually all day and the Flex will continue to work, despite showing the “empty” battery level on the iPhone app.

Enjoy these posts? – Tell a friend

Recommending readers is one of the highest compliments you can pay to a writer. If you enjoy what you read here, or you find the posts useful, tell a friend! Find me online here:

Twitter | Facebook | Google+ | Blog | RSS

Or use one of the share buttons below. Thanks for reading!

How I Automatically Capture Driving Data From my @automatic Link in a Google Spreadsheet

I have been using the Automatic Link in my Kia Sorento since December. It is a good little device that plugs into your car’s data port and pulls out all kinds of interesting information about your driving habits. For a while, you needed the iPhone app to browse the data, and the data itself was not extractable in any easy way, but no longer.

A while back, the Automatic tracker became available on IFTTT, with a bunch of triggers that can be used in automation workflow. One of those triggers is when a new trip is completed. So I created a recipe in IFTTT that logs the data of each completed trip to a Google Spreadsheet. For now, it logs all of the data, even though I might not use all of it. The data is logged within 15 minutes of completing a “trip” (going from point a to point b and shutting of the engine). Here is a list of the data that gets collected in the spreadsheet:

  • Car
  • Start Time
  • End Time
  • Duration
  • Distance (miles)
  • Average MPG
  • Fuel volume consumed (gal)
  • Fuel cost (dollars)
  • Hard brake count1
  • Hard accel count2
  • Duration over 70 MPH (minutes)
  • Duration over 75 MPH (minutes)
  • Duration over 80 MPH (minutes)
  • Trip Map URL
  • Start Location Longitude
  • Start Location Latitude
  • Start Location Map URL
  • End Location Longitude
  • End Location Latitude
  • End Location Map URL

The spreadsheet looks something like this:

Automatic Link

The great thing about this is that, like the FitBit Flex or my Google Writing Tracker scripts, the data is collected automatically. This is, in my opinion, of critical importance for personal analytics, because any time you have to take for manual actions lessens the likelihood you’ll continue to collect the data. For this data, all I have to do is drive.

I only have a week of the data so far, but it has already confirmed what we already knew: we have an incredibly good commute to and from work. I live about 5 miles from the office (5.18 miles on the roads according to the Automatic Link). When we leave the house at 7:16 am (as we did yesterday), we arrive at my office at 7:28 for a total trip time of 13 minutes. (Kelly has to then catch the Yellow Line from my office to her office in the District.) Coming home. Our reverse commute in the evening takes 12 minutes, despite being right in the middle of rush hour.

There are a few things I am trying to tweak with the spreadsheet. One downside is that the data/time is entered as a text field instead of an actual date/time and that makes some charting difficult, but I’m working on some code that will convert this automatically. Then, once I have more data, producing some charts and plots similar to what I’ve done for writing and walking should be easy.

One thing I’ve learned from this that I’d never thought much about before is the cost of our commuting into the office. Looking at the fuel consumption of our commute and Automatic’s estimated fuel costs, our commute costs us $1.85/day. That amounts to $9.25/week, or assuming we work 48 weeks out of the year, $444 in fuel costs commuting to-and-from work each year.

That number is actually high because there are days when we both work from home, but I suppose the number wouldn’t be less than $400/year.

I’m looking forward to delving deeper into this data once I have more of it to make it more meaningful.

ETA: I’ve embedded my IFTTT recipe for this automation below, for easier access.

IFTTT Recipe: Export Automatic Trip Data to aGoogle Spreadsheet connects automatic to google-drive

  1. The tracker detects when you brake too hard as part of its system for analyzing fuel consumption performance.
  2. The tracker detects when you accelerate at a rate that burns fuel in a less-than-optimal way.

Writing By the Week

Yesterday, I mentioned how I completed my first week on the second draft of my novel, and provided the number of words I produced for the week. It occurred to me that in the past, I’ve been reporting mostly daily word counts and trends, and never weekly, so this morning, I took a quick look at my data to see how things look at the week-level as opposed to the day level. Here is a plot of the words counts for every week I’ve been writing, going back to the week ending March 3, 2013:

Word Counts by Week

The red line represents my average over the last 59 weeks, a number that comes to about 5,800 words/week. As you can see, last week, I broke my overall average for the first time since back in January. Those 10-, 12, and 13,000 word weeks you see last summer was back when I was racing to finish the first draft of the novel.

I took one more quick slice of the data, plotting it as a histogram, and came up with the following:

Word Per Week Histogram

From this, you can see that I’ve written between 4,500 – 6,000 words on 21 different weeks. I’ve written between 6,000 and 7,500 words on 15 different weeks. There is only a single week where I’ve written less than 3,000 words, and three separate weeks, I’ve written more than 10,500 words.

There are some hard limits here. As I’ve mention before, my biggest obstruction is not a lack of ideas or will, but time. I only get about 45 minutes/day to write, on average. If I could slowly work that 45 minutes up to an average of 80 minutes, I could come close to 2,000 words/day (14,000 words/week). But it will take time to get there.

The Elusive 10,000 Hours

A week or so ago, I calculated how much time I spend writing each day, based on 400 days worth of data that I’d collected1. The data showed that I average about 42 minutes and 15 seconds of writing time per day. I have written every day for the last 267 days, and I’ve only missed 2 days out of the last 410. So I am writing every day.

This is the first period of time in my life where this has been true, although I have been writing stories now for more than 20 years. Prior to 2013, I wrote in small, dense scraps of time, producing one, or maybe two stories a year, and spending perhaps a total of 20 hours writing for the entire year. However, from February 27, 2013 to the present moment, I’ve spend about 289 hours of my life writing.

In the big picture, it is not all that much. In that same span of time, I’ve spent about 2,400 hours of my life at the day job. I’ve spent approximately 2,700 hours sleeping. The time I have for writing is roughly 1/10th the time I spend at my day job and 1/11th the time I spent sleeping.

I was thinking about this in the context of the 10,000 hours that it supposedly takes to become an expert at something. 10,000 hours sounds like a lot, but in practice, it really isn’t. If you could work at something–say, writing–for 8 hours a day, you’d hit your 10,000 hours in less than 5 years. Even half time, you’d still hit your 10,000 hours within a decade.

But I don’t have that kind of time. I can spend, on average, about 45 minutes/day on my writing. Occasionally, I can spend more time, but that is offset by the days that I spend less time.  Which means that even writing every day of the year, which I do, I’m spending less than 300 hours a year writing. It’s not difficult to take that number and figure out how long before I hit my 10,000 hours. Excluding everything that came before last February as marginal (I’ve written more in the last 400 days than in the last 20 years put together), it will, at my present pace, take me nearly thirty-eight more years to hit that magic 10,000 hours. At which time, I will be 80 years old.

This assumes, of course, a steady-state, and that is unlikely. I try to set goals that are obtainable when it comes to my writing, things that are in my control. Rather than have a goal to sell 10 stories or win some award, I pick things like: average 1.5 hours of writing per day by the end of 2015. Where will the time come from? That is part of the challenge. I don’t want to take away from family time. That leaves other areas of my life. I’m not giving up my day job, and I’ve more or less optimized my sleep. That leaves little wiggle room.

So I’ve started to prioritize what’s important. I started the second draft of my novel a few days ago, and I recently gave up my book review column at InterGalactic Medicine Show, and yes, the two are related. This was a paying writing gig, but the time it took can be redistributed toward my writing. Then, too, I am constantly on the lookout for ways I can automate things so that I don’t need to spend time on them.

I try to keep my goals modest, but I would love it if I could reach the stage where I could write 2 hours per day. That’s 2.9 times what I currently manage, and that means cutting my time to hit 10,000 hours from 38 more years down to 13 years. Now we’re talking. I’ll be around 55 years old and much closer to thinking about retirement from the day job. And with the practice and expertise I will have gathered from 10,000 hours of writing, who knows, maybe at that point, I’ll be able to support the family with my writing.

  1. 412 as of today.