Feltron Post-Mortem

a/k/a My Qualified Disaster
a/k/a The Trouble With Tech

previously on dy/dan

We started with four variables (text messages, beers per day, etc.) which we tracked for 2.5 months in quad-ruled notebooks attempting to transform the quotidian details of our lives into extraordinary infodesigns a lá Nicholas Felton.

This was a departure for me. A tech-driven, student-led, design-infused mathematical project. Things went wrong.

This is a comprehensive autopsy of our Feltron Project. I post it here, in its entirety, a) for my own review next year, b) for your criticism. If you aren’t in the mood for the full, bone-by-bone dissection, please scan down to the section headed What Really Happened. These are problems I don’t know how to solve.

The Lesson Plan

a/k/a What Was Supposed To Happen

  1. We selected variables.
  2. We discussed them, making them more interesting (disaggregating “hugs per day” into “boy hugs” and “girl hugs”) and more manageable (tracking “fast food I eat” instead of “what I eat”).
  3. We tracked them for ten weeks, checking ourselves for consistency every two weeks, and then we stopped.
  4. We spent one hour marveling over Nicholas Felton’s annual report, dissecting it for meaning, identifying the mathematical operations (average, maximum, minimum, sum) and the mathematical forms (pie chart, line graph, histogram, stacked bar graph, map) he used.
  5. We spent six hours entering our data into Excel sheets.
  6. We spent two hours teaching and deriving ten facts of our lives using average, maximum, minimum, and sum functions in Excel.
  7. We spent two hours teaching and deriving four graphs of our lives using pies, lines, and bars.
  8. Raw facts and graphs in hand, we spent thirty minutes discussing and distilling Felton’s graphic design savvy into the two principles I thought my freshmen could reproduce with crayons and paper if they had nothing else:
    1. colors, Felton uses a two-color design (shades of black, shades of blue) which, apart from distinguishing his hierarchy (titles in black, data in dark blue, accents in light blue, etc.) keeps down costs when designing for a large print run.
    2. grids, the kind your eyes can’t see but which your brain loves, the kind which imposes order on what would otherwise be a completely disordered data set, so while Felton jumps from music to movies to drinks you know where to find everything.
  9. We spent another two hours in class tying up loose ends in Excel and then a week designing our Feltron Projects.

What Really Happened

a/k/a Help.

  1. Only 55% of my students submitted the final Feltron ProjectControlling for age: 48% of freshmen and 63% of upperclassmen completed the project..
  2. Many of the other 45% stopped tracking early in the project, which meant assigning them review work, new work, or busy work while everyone else worked in Excel.
  3. Those who kept up with the project quickly staggered their progress (based on pre-existing computer ability, typing speed, and attendance) which saw me dashing between desks, explaining and re-explaining the same procedures over and over again.
  4. Our mobile computer lab a) comprised just fifteen laptops, and b) was available for check-out only once a week, c) if that.
  5. Kids lost work. I had them send their Excel files to themselves and then download the attachment the next day. Trouble was kids sent old files to themselves or they named files computer arsenic like “<<xxxx….davidsfeltronz!!!….xxxx.xls>>” which put both Excel and Gmail into simultaneous cardiac arrestFor the record, I originally sought GoogleDocs out for this project but they maxed out at something like fifty rows where we needed hundreds..
  6. I overestimated my students’ computer fluency. Name it: locating saved files, opening programs, using a trackpad, using modifier keys, sending e-mail. These tasks all required constant, patient re-explanation. Missed that mark by a country mileThere were exceptions, naturally, but Digital Immigrants™ outnumbered Natives™ at 15:1, many of which Natives one day, I have little doubt, will grow up to be edubloggers..
  7. None of them had used Excel before. Ever. Many didn’t have it at home. One triumph of this project – recognized by a lot of students – is that my kids are now somewhere in the top quintile of Excel users. This will doubtlessly prove useful again in their lives – not in the when-will-we-ever-use-this-in-real-life? sense, like they won’t be able to find food or shelter without Excel, just that it will open up a lot of interesting opportunities.

What Mattered

a/k/a Grading

  1. Faithful Tracking
  2. Interesting Findings
  3. Clear Design

Students ranked themselves on a ten-point scale across each index. Given how deeply we had immersed ourselves in exemplary work over two-and-a-half months, with only a few exceptions, I gave them exactly the grades they felt they deserved.

What I’ll Do Next Time

a/k/a If There Is A Next Time, Obviously

  1. Host screencasts online demonstrating essential Excel proceduresincl: sorting columns, using formulas (avg, min, max, sum, countif), saving/sending work, creating new sheets, filling down the date..
  2. Strengthen our analysis. A student’s text message graph plunged for a week when her parents confiscated her phone and spiked when she pulled a boyfriend in May. Students positively thrilled to see those connections but we didn’t build any of that analysis into the project grading. Should’ve.
  3. Employ a Kuropatwa-esque rubric to better inform kids what constitutes “clear design” or “faithful tracking.”
  4. Discuss design in greater depth, incl.
    1. showing them what my own Feltron would look like with rangy, mean grids or spasmodic colors;
    2. showing off the good and bad from this year’s class;
    3. comparing/constrating Khoi Vinh’s approach to grids and David Carson’s insane anti-grids;
    4. showing them Aesthetic Apparatus’ beautiful work in just three-or-fewer colors;
    5. compare 3D graphs alongside 2D hoping a lot of students will reconsider the choices they’ve made in life.
  5. Make a more obvious point of my own Feltron Project. Playing along with your students isn’t even optional here. I made sure I ran through the collection process with my students (for empathy, if nothing else) but I should’ve made a larger point of my own struggle and process.
  6. Find collaborators. This was insane. I should not have gone at this aloneAny takers?.

Students On Feltron

Just do a month.

JG, smart; we’ll multiply a month by 12 to extrapolate for a year.

Everyone should track the same thing because it’d be really cool to see which people are like you.

BP, also smart; resolved, then, that we’ll select three variables independently of the class and then select a common classroom variable for the fourth.

I like the chalang. It feels like I acopolished something hard and it made me feel good.

BS, sic sic sic; whose mother, in an IEP meeting, said of his Feltron notebook, “He carries it everywhere.”

Felton On Feltron

Nicholas Felton consented to an e-mail interview on his process which will appear in this space tomorrow.


I have installed student work – everything from awful to exemplary, but mostly exemplary – into a Flickr set.


  1. Feltron Project Outline
  2. Nicholas Felton Analysis Sheet
  3. Excel Chart Illustrations
  4. Excel Formula Sheet
  5. Map Infograph Template
  6. Final Review Sheet

To Conclude

This was a different, necessary kind of insanity for me to finish my fourth year teaching even a little eager for a fifth. The price tag was steep. To accommodate this time-sucking project-based learning, we skipped a third of our logic unit in Geometry and fully jettisoned last year’s Platonic Solids project.

If I weren’t already guzzling away at this barrel of standards-based Kool-Aid, I’d write something agitated and truly inexcusable here about curriculum narrowing or the time cost of NCLB, but I remain convinced we need to settle on a list of necessary skills and then decide horse-in-front-of-cart-style on the best tools and projects to teach themNoted here: Jay Greene’s j’accuse directed at teachers who complain that NCLB exigencies leave them with no time for fun project but who also wile away the last month of school with parties, assorted time wasters, etc. We didn’t start computer lab work with Feltron until after our round of state assessment.. I do not know if this was that.

There are twenty-four hours. No exceptions. I’m uncertain Feltron was the best use of our time.

I put Feltron to rest now, surely the weirdest assignment I’ve concocted in a four-year career. I post this here to solicit the usual gallery of critique and construction but also because, at some point in this whole blogging thing, I forgot how else to end a project if not with rigorous and public self-critique.

I'm Dan and this is my blog. I'm a former high school math teacher and current head of teaching at Desmos. He / him. More here.


  1. Thanks for this. I had been trying to decide what to do with my stats project next year, this will help. I do things a little differently, but I don’t know if you would like it.

    Data Collection takes one day. I give every kid a packet with 8 sets of 3 questions (qualitative, discrete, and continuous data), and every kid answers them all in one class period. I trade buy-in (kids not choosing what to track) for access (everyone submits data and gets a pile of data to work on). And as math projects go, buy-in isn’t bad. They like to see how they compare to their classmates.

    Overnight I take both classes’ data and chop the 8 sets of questions up with a paper cutter. In class, they get in groups of three or four, and each group gets a data set – everyone’s responses to 3 questions.

    I limit their options for what types of graphs they can choose (histogram, box&whisker, circle graph, scatter plot) to align with what the NY standards expect, but they make the decisions about which types of graphs to apply to their three questions. I give the option of using computers or creating a poster. I give each group a big packet of Excel instructions for each type of graph. This cuts down on my having to run around to groups like a crazy chicken. They also have to calculate measures of central tendency for their numerical data and answer-in-complete-sentences a few analysis questions for each.

    All this group work takes around 5 class meetings (I have 43-minute periods, not blocks). The biggest challenge I have is helping them organize and tabulate their data so they can work with it easily. They are assessed by giving a presentation to their class and invited dignitaries and graded with a rubric. The computer work makes this nice, we just throw their stuff up with the projector. The posters are more awkward.

    I like this project, but I’m not happy with their retention of certain details yet. Later today I will post my related documents at f(t) if anyone is interested. Right now I have to take my visiting mom for breakfast and enjoy the last non-work day of summer!

  2. Thanks for posting all these details! Very much looking forward to reading it more carefully. Getting more authentic work into math class while still covering the college prep standards just isn’t a trivial task, and this is a particularly interesting attempt.

  3. Still slogging my way through the Rhee post (and I’m old, but just getting certified, and…well…I’m more like the old’uns than the young’uns on this one. A decade or two of life experience and seeing the politics of situations that shouldn’t have them and one develops a fondness for health benefits and places wherein bad people can be removed…but that ALSO protect the good and best people.)

    Quick thoughts here:

    Shorter data collection, especially since it’s four things. Either cut the number in half (one individual, one for the whole class) or cut the time to a month.

    Data set(s) to provide to those who drop out. Enables/requires them to complete the task.

    Did you keep track of the excel problems? As mentioned above, develop written (or websited) procedures to hand out in an FAQ sort of format.

    I think your immigrant to native ratio is right on. Kids’ computer skills are *not* that great. I agree that you did them a big favor pushing them through this (sounds like play-doh extrusion when I say it like that).

    I heartily agree on the wasted time.

    However, my kids are in a district that is going to look at timing elementary curriculum delivery to the minute in a couple of schools this year — 4 minutes on pre-story vocabulary, x minutes reading story, z minutes answering y questions. I fear for the ability to tailor anything, to do anything that’s useful and educational, and fit it into wasted moments (or hours provided by not having wasted moments).

    We already have every 6 days testing (even in Kindergarten, where flipping the pages correctly is harder than the test itself and where the term “playscape” was used about a playground. Way to reach the 5 year olds, wondering where the heck that might be when all they could see was a playground. :-p ) It’s fine for assessing what happened in the previous 5 days (though many teachers could give you the results from their observations of the kids and their work). However, the next assessed unit begins immediately. If you find out your kids need more work…well, too damn bad! Try to fit it in the cracks and definitely make sure they don’t screw up on the next test.

    I’ll be student teaching this starting…really soon! Then I’ll really know if there’s lots of wasted time or any time for enrichment at all.

  4. I need an after-standardized-test-over-project for my AIMS Math class. This might be worth a spin.

    (I would also a.) do some parts off-computer b.) follow JG’s suggestion on time spent given I wouldn’t have more and c.) work on mixing an component dealing with either finances or current news stories.)

  5. While I understand very little about this “maths thing” you blog about, I love that you share your failures as well as successes. Not that this was a failure but you’ve identified weaknesses and considered next steps.

  6. I find that when I’m not fluent in a program but also not completely lost, a printed (or PDF…) reference helps and a screencast takes more time to watch than it’s worth.

    Another thing that would make this task go faster would be some contact with Excel earlier in the class, with a project that was designed to keep everyone closer together in the process. Of course, then you need to find the time for that project.

    A data set for people who lose their data is pretty much invaluable, too. (As Jen says.) Even if everybody was a perfect tracker, what are you gonna do with the kid who just transferred into the school district?

  7. Dan –

    This is a powerful post-mortem. It’s serendipitous that you posted this since I was about to go through your archives to find the details of this project.

    I’m teaching a class this year (long story) of IB math students and they are required to do an “internal assessment” (read: project) as part of their diploma requirements. I was hoping that Felton/Feltron might be just the thing.

    Thanks again!

  8. The one detail in this kind of data presentation that I obsess most about is the units used for expressing the data. Good units can have a similar effect to the visual design, like nice colors even grids.

    For example, saying you at 800 pieces of gum over the course of the project sounds much different than saying you ate 5 pieces of gum per day during the project. They convey different things. Often times absolute units (800 pieces of gum) confuse me because it makes everything seem huge, whereas normalized units are often easier to visualize (5 pieces of gum per day).

    A similar comparison can be made for electric power units: A 100 Watt light bulb uses 3.2 GJ in one year. 3.2 GJ sounds like a lot, and I don’t really have an intuitive grasp for it. 100 Watts is equivalent and makes more sense to me.

  9. Dan, Nicholas has released a website called daytum.com and once it is operational your students would be free of all the “computer skills” issues. They could just maintain their stats on daytum. Worth checking out.

  10. Dan,
    Things to think about:
    1. Start smaller to teach the Excel skills (like sorting M&Ms out of a package) to make sure everybody has the same skill set.
    2. Could they have had the option to work in pairs? Or could each freshman pair be in charge of only one data set? and them compile into a group project. Older kids could do “4 slide project”
    3. To get the data could different pairs bring different data to the group, ie somebody do food, some text messaging, some TV watching, some Facebooking, or whatever— ( like Jen suggested)
    4. Could you have a graphic designer come in and talk about the design elements?
    5. Open a ning or Moodle to storing all student work (after discussing file naming)!!
    6. Offer to do something with his/her students while the business apps teacher teaches Excel to your kids.
    7. Assign an online Excel tutorial as homework, there are tons online.
    8. Write a grant for 1:1 computers in your classroom, the math connection would be a huge draw since I would think few people write tech grants with a math focus. Let me know and I’ll send you info on which grants to apply for.

    OK< that was fun. N

  11. A dummy data set seems like a good course (certainly better than nothing) but it was nice that no one could copy anybody else’s work. I wonder if I could randomize the variables (eg. the number of soda a fictitious person drank in a day) in Excel. Probably wouldn’t be too rough.

    @Steve, true, and then there are the times when the impact of a gross number is fun, necessary, and interesting, like (eg.) “I have consumed enough caffeine in the year-to-date to kill three bull elephants” instead of “I have consumed three grams of caffeine each week.”

    @Tim, I saw Daytum a coupla days ago and it looks pretty and useful. I worry, just as I did with Wordle and just as I do with basically every Web 2.0 tool, about what it makes easier. I need to simplify the collection process, which Daytum does, but I also want them to get their hands dirty creating the infographics, which Daytum, unfortunately, also simplifies.

    I’m not sure what Daytum does to academic rigor, in other words. Where’s my beta invite?

    @Nancy, good suggestions. Can you e-mail me some grant info? dan at mrmeyer dot com. Thanks a mil.

  12. The 2 things I never got with the Feltron project were taste and time.

    The design (Feltron’s) didn’t do it for me. It wasn’t that I hated it, and I could see it was accomplished, but I burrowed down into it and it never resolved into anything for me. Just stayed as graphs and grids and numbers in different colours – not to my taste. I would never have been able to digest this as a child.

    Time/effort is obvious – 4 variables per day for 10 weeks is a pretty heavy obligation, and it would have sparked some emotional baggage/resistance in me because I would have taken it as “what I did on my holidays” to the power N – and I never accepted having to document myself on demand.

    I can see why you’re going with it, and the Excel experience is surely useful but I really doubt that you can refine it to get over the 2 points above: trying to inculcate a design sensibility + forcing self documentation + teaching math, is too much.

  13. I haven’t read the comments, so please forgive me if I am repeating something that someone else said.

    Aside from screencasts to cover important excel concepts, also consider assigning the strong students to the weak students to help them out. Only don’t make it sound like “assigning”.

    Try using Google documents. They have a spreadsheet. Save automatically. Are hosted on the web. Cannot be killed by computer arsenic file names. And create graphs and charts. Oh, and did I mention that they are free? docs.google.com

    When doing this, have students generate a hypothesis. This is the science teacher in me speaking. But I find that when students are asked to deal in data, it helps to give them a “so why am I wading through this data”. Here’s how I would do it:

    Give each student an envelope. Have them write down their predictions on different variables. Example: “I believe that I spend more time studying than watching TV.”

    Put those predictions inside the envelope. Store it in the classroom. Tell students that you are going to check back at the end to see how many really know themselves.

    I like the idea of keeping data every day for an entire month. That give 30 (or perhaps 31) days of data (provided it is not February) on four different data points. That’s a lot of data for high schoolers. Heck, that’s a lot of data for college kids, too. If they master this, then they can move upward and onward.

    Not knowing your student’s home computer/internet situation, I’d make an offer to have the students go ahead and record the data right into their spreadsheet of choice. Then the ones who come in on data!entry day 1 and claim to be, “All done,” can easily be reassigned the role of spreadsheet expert and sent forth to enlighten their fellow classmates.

    I’d also add an analysis piece. First I’d have them compare their actual results to their earlier predictions. Then I’d have them try to come up with some reasons why the two didn’t match. Why, for example, was their TV time higher than their studying time? Why did they feel otherwise? Where is that disconnect coming from. And how does it relate to their grades in school?

    I think this is a great assignment, and certainly did better than expected giving that it was the first year you were trying it out. The question to ask when doing it again is how to make it more authentic; how to turn this assignment in to answering a question the students really want to buy in to answer. It appears that you love simple visual representations of data, but these students are young yet and won’t do it for such an abstract concept. Give them something more concrete.

  14. I’m trying to get my class (sixth graders in US-speak) to do a bit of basic info design using maths from the Beijing Olympics where (hopefully) the data isn’t too hard to find, allows them some choice and gets them thinking about the best choice of graph or numbers. Like myself, their ambition exceeds their actual ability so getting them to think simple has been the key to producing something that is useful. For example, one student compared the differences between heights of the women’s gold medal winners of high jump and pole vault – but even getting to choose wisely (column graph is a better visual for height than a bar graph) needs some timely counsel. Plus eleven year olds think that no slide is complete without images, 3-d graphs, Word Art titles etc. – getting the “less is more” idea is quite a challenge. But I do think that they do gain an appreciation for mathematical comparison especially when they see how little separates gold from silver from bronze or how much higher an athlete can levitate using a springy pole – and it’s the appropriate graphic design that makes this clear.

  15. Can only hope I can get to this degree of reflection in my own work this year. This was great.

    What I do not see, however, is a disaggregated treatment of your successes in this project. Not doing so may be a function of attempting rigor, but as in any analysis, not overtly acknowledging what went well can be just as misleading to the analyst as glossing over the mistakes.

    Here’s what I would ask in pursuit of ferreting out success:

    1) Of the 55% of kids who turned in a project, how many of them demonstrably mastered– or at least made significant progress towards– the standards addressed?

    2) How many of them mastered some other set of tangible skills or concepts that are not specifically standards-based, but still valuable enough to spend your time on? (Ex: Excel spreadsheets, not naming files like pornographic films)

    3) What *intangible*, qualitative successes did you have?
    (I’d say student BS’ comment was a prime example.)

    And the punchline: Are your answers to these three questions enough to justify trying the project again?

    P.S. You have to try it again anyway. No self-respecting scientist I know is happy with only one data set.

  16. I teach an intro to stats class and read your project with interest. I have done similar projects of those described here, but with a difference: I try to have the students survey and analyze a topic that is interesting and meaningful to the school. For example, we recently completed a School Safety Survey whereby each student surveyed 5 students each from different grades, we compiled the data and students analyzed it individually in Excel. We were able to identify key problems at our school (for the whole school and broken into middle and high school). When we have learned all graphing techniques, we will be posting the results on poster graphs for all to see and it will be included in our newspaper. Our faculty and admin were also interested to see the results.

    Second, I have found conducting an individual project on a topic of the student’s choosing successful. I assign weekly parts of the project with drafts and final versions due. There must be 2 variables to investigate and the data must be readily available from a reliable source. Analysis is through Excel. This culminates in a final presentation with appropriate graphs and meaningful summaries of findings.

    Each year I have been able to refine these lessons, striving for hands-on or real-life applications and ideally, interdisciplinary. I find there is so much to fit into my semester course it is difficult to parse things down! I feel satisfied knowing that this is the only class where students come out with a very real and applicable knowledge of Excel and stats that they utilize in other/upper classes.

  17. I know I am getting late into the discussion but a few comments (or obvious observations) I would like to agree with that Dan said.

    #1) students computer skills not as high as we think they are. Kids can text message and use cell phones and ipods, but skills like saving files and changing directories and Excel aren’t obvious. When I started teaching I thought what Dan thought (that these are things that the kids will know coming into your class), but I still have to go through them year after year.

    #2) 45% handing it in: one of the potential difficulties with this project is the fact that students have to track their activities for two months to get the data. I realize the purpose was to give the students ownership over their work and to make them feel that this was their project (as undoubtedly took place for those that submitted it), but tracking your daily life is not as easy thing.

    #3) You took a risk and it didn’t work out as much as you wanted to. You know what Dan, most people wouldn’t be able (or confident enough or whatever) to take a month risk on a project. Since you truly believed in it, I know that you will have done lots of reflections on how to improve it next time (if there is a next time as you say).

    As usual, your blog inspires (even when things don’t go your way)

  18. Well, I realise I’m getting in very late here, but this really blows me away. Partly because, once again Dan, you have written about school in an amazingly eloquent and insightful way. Thanks. Also because, if anyone did this around here (and wrote about it so well), they would be fêted as a hero and invited around to talk about their project, maybe even to give courses. You, on the other hand, have simply analysed, evaluated, learned and moved on.

    Why the difference? Are you folks in California so focussed on ‘teaching to the test’ that there isn’t space for messy, creative, hands-on learning? Or is Scandinavia (where I am sitting) so focussed on positive learning experience that we have forgotten our responsibility for teaching long lists of hard skills? I don’t know, but the contrast amazes me.

  19. Dan,
    Which class did you do this in? It seems like more than one class would be too much, but fortunately, you don’t know those words.

    I was thinking this could possibly an appropriate end of the year assignment/project.

    How did you make the graphs for days out of the month that you blogged? Are those from Excel?

  20. I’m pretty sure I ran that graph through Adobe Photoshop.

    This project is fun for classes young and old, though you may decide to vary the constraints a little for the younger set. I did this in all of mine.