Toggl: In Detail

In which I find out what I do all the time...again...

N.B.: I've written this post a bit later than intended. I now have a new job as a developer and will post about that after about 6 months.

Following on from my last post I delve deeper into the types of work I was doing in my previous job and my side projects.

MR Solutions

Projects

Percentage Time Spent on Projects at MRS

Figure 1: Percentage time spent on each project at MRS. 62.12% of my time was spent on customer sites (either physically or remotely); 16.79% was spent on Administrative tasks; 15.35% on coding pulse sequences (my job title); 2.59% on training (customers or distributors).

Fig. 1 shows the percentage time I spent on different projects. I grouped anything related to customer sites (remote or on-site) into a single project as they are similar tasks. Nearly 2/3rds of my time was spent on these customer focused tasks.

In second place, I spent 16.79% of my time on administrative tasks. I'm not yet sure how much time I would expect to go on admin. I'll have to compare this to my new job to see if it has reduced. I know that a significant portion was waiting on emails when I was on support for the day. In general I would aim to reduce the amount of time wasted on emails. 

In third place, with 15.35%, was Pulse Sequences. Considering my job title was "Pulse Sequence Developer" this is a little disappointing but not surprising as I was often travelling to sites instead of writing and testing sequences. As it was a small company most people were working outside their original job description.

The 2.59% on training is probably an anomaly as I helped organise and run a week of distributor training during this period.

Tags

Figure 2: Word cloud of most used tags.

Fig. 2 is a tag cloud weighted by time spent on each tag. I'm going to ignore the OS ones as the majority of my work was on Windows due to the fact our software ran on Windows. Table 1 contains the percentage time spent by tag. N.B: These add up to more than 100% as you can use multiple tags per time tracking event.

TagPercentage Time Spent / %
Pulse Sequence 26.43
Travel 23.91
TeamViewer 21.84
Support 20.06
Remote 18.63
Physics 18.45
Coding 11.36
Install 10.94
Hardware 8.06
C/C++ 5.11
Training 4.62
Updates 3.27
Meeting 1.66
Documentation 1.66
Tickets 1.27
SQL 1.00
Python 0.48
Research 0.46
Report 0.28

Table 1: Percentage time spent on each tag. As multiple tags can be used per event the total is more than 100%.

So based on tags I spent more time on Pulse Sequences than on travel. This is because there was often Pulse Sequence related work for customers such as optimising parameters or training on how to use a new sequence. Therefore this tag covers more than the coding specific project mentioned above. 

The Travel tag covers any form of travel related activity such as train/plane/taxi/ferry or waiting at an airport. It wasn't applied to any customer activities when I was on site. You can see that I spent nearly a quarter of my time travelling. This is something I delved into a bit more in the previous post.

TeamViewer, Support and Remote tags are all similar in percentage (around 20%). A small amount of TeamViewer usage outside of direct support may explain the slight difference though I'm not ruling out the possibility that I mistagged a few support requests in my hurry to help the customer.

It's good to see that in my role as 'physicist for hire' I at least managed to clock up 18.45% of my time on physics related activities. This included sequence design and data analysis that made use of my knowledge of physics. This was generally the work I most enjoyed doing (alongside pure coding). At heart I love to solve a good problem.

Coding comes in at 11.36%. The disparity between this and the Pulse Sequence project is due to design and testing work that didn't involve direct coding. The fact that C, C++ and Python don't add up to this percentage is because I only started to track specific coding languages about halfway through this 20 week time period. There was also some SQL work related to our ticketing database that contributes another 1%.

The rest of the tags mostly pertain to admin work and maintaining customer installations.

Side Project: Podcasting

Percentage Time Spent on Podcasting Activities

Figure 3: Percentage time spent on each podcasting activity. 66.89% of time is spent on editing, 21.20% on recording, 5.12% on design work (such as logos) and 2.19% on script writing.

Fig. 3 shows the breakdown of podcasting related activities. With podcasts the majority of my time is spent editing and that shouldn't come as much of a surprise. I think the most time I have spent editing a single episode was around 14 hours but that was due to specific issues with a lost recording*. The overall percentage of editing should come down over time as I pick up tricks and just get better at it overall. Though I still expect it will be longer than the recording time in future.

Again it isn't a surprise that the second largest chunk is recording the episodes. HOMT episodes are often 90-120 minute recordings and that can be extended if we have issues with our VOIP setup. Dr Wilko's episodes still take around 30 minutes to record despite only ending up in 4-5 minute episodes when the editing is done†.

Design work has taken up 5.12% of my time so far. I expect this will drop next time I analyse my tracking data as most of that was upfront work such as creating show logos.

Finally Writing takes up 2.19%. As I only write scripts for Dr Wilko's it isn't surprising this is such a small percentage. I also tend to write them in breaks when I have the time so it is possible there might be some under reporting there.

*Kudos to Will for spending a stupid amount of time re-recording his parts into the gaps we left.

†See below but the overhead of making a cocktail really adds to the recording time.

Pecentage Split Between Podcasts

Figure 4: The percentage split between the two podcasts. HOMT takes up 69.88% and Dr Wilko's takes up the other 30.12%. Dr Wilko's has twice as many episodes but they are shorter and require more planning.

Fig. 4 shows the percentage split between the 2 podcasts. Based on episode length you might not expect this even with the weekly release schedule of Dr Wilko's compared to the roughly fortnightly release of HOMT. Dr Wilko episodes have much more overhead though with script writing, recording as I make a cocktail (sometimes requiring multiple takes) and a more complex edit (most of the time).

Tom Out!