Analyzing a database of over 6 million online sales of residential HVAC filters from 2008 to 2017

Back in May 2018 I was invited to give a talk about our research on mass-based HVAC particle filtration metrics at the National Air Filtration Association’s (NAFA) annual Tech Seminar, and I made a pitch to the audience of mostly filter manufacturers and retailers that the academic community needs more information on the residential air filtration market. Basically: how can we estimate what fraction of the U.S. population uses what kinds of filters? I’ve tried to get at this question before. In a 2014 paper in which we needed some statistics on the national distribution of filter ownership, we really had to make some back of the envelope estimates from a variety of fairly low quality sources (including some personal communication with a colleague in the industry who gave us the thumbs up on our estimates, saying “that sounds about right”).

Fortunately, one of the audience members at my NAFA Tech Seminar presentation, Rich Scott from Discount Filters, a major online retailer of HVAC filters, volunteered to provide me some sales data. Rich and his team provided me a database of over 6 million filters they have sold to (mostly) residential customers from 2008 to 2017. I say “mostly” because they sorted the data by commercial vs. residential address, although they know that not to be the most accurate description of their customers, but sufficient enough for this purpose. There are a few other caveats worth noting about their data as well, including: 1) the sample is large, but might not represent the entire U.S. residential marketplace (e.g., online consumers might differ from in-store consumers; some of their customers are large management organizations that buy lots of filters, which can dwarf individual consumers); 2) they don’t sell every type of filter under the sun, so the data are obviously skewed to the products they sell (e.g., they only sell MERV filters and not other metrics like FPR or MPR; these are mostly 1-inch filters and deeper-bed 4- and 5-inch filters combined into one data set, I don’t think there are any 2″ filters here); and 3) some of the trends over time capture more than just changes in sales due to consumer preference alone (e.g. changes are also affected by the marketing efforts and product offerings of this particular retailer, which have also changed over time). Nevertheless, let’s look at the data! Many thanks to Rich Scott and Tyler Dishman at Discount Filters for providing these data and being so willing to share the data publicly. 

First, let’s look at the data set of sales over time. Their sales grew from ~180,000 filters in 2008 to ~1.3 million filters in 2017. The total number of filters in the data set is about 6.4 million from 2008 to 2017.

Total number of (mostly) residential filters sold by from 2008 to 2017

That’s a lot of filters! Next, let’s look at the trends of fraction of total sales of the different filter types that this retailer sells over the same time period. 

Fraction of different types of (mostly) residential filters sold by from 2008 to 2017

You can see MERV 13 steadily on the rise since 2008, while the fraction of MERV 8, 11, and fiberglass (i.e. low MERV, probably less than MERV 4) filters are all fairly steady over time, albeit with some year to year variation. MERV 10 filters have been decreasing but the retailer mentioned to me that has less to do with the filters themselves and more to do with production and marketing (in other words, the retailer has moved away from MERV 10 in favor of some of the other filters). Washable, MERV 16, and activated carbon filter sales were steadily low throughout the period.

Last, let’s look at a snapshot of these same data in 2017, the most recent full year of sales in the data set. 

Fraction of different types of (mostly) residential filters sold by in 2017

In 2017, for this particular retailer, MERV 11 filters made up the largest share of sales at about 35%, followed by MERV 8 at about 29% of sales. Fiberglass filters were third at around 16% followed by MERV 13 filters at about 14% of sales. In fact, fewer than 15% of filters sold were MERV 13 or higher. These data suggest that at least for shoppers on this major online retail site, not that many people are using high efficiency filters in U.S. residences! I hope you find these data helpful.

Some thoughts on conducting energy and environmental research during the first 2 months of the Trump administration

I don’t think of Twitter threads as a particularly helpful means of communication, but it’s a good way to rant about my experience conducting energy and environmental research during the first 2 months of the Trump administration:

Thanks to Chuck Haas at Drexel University as well for making a Storify of this thread.

New batch of 3D printer filament emissions tests available

We have released 7 new reports on ultrafine particle (UFP) and volatile organic compound (VOC) emissions tests conducted in our lab in 2016 using a Lulzbot Mini 3D printer and 7 different commercially available filaments. Scroll down to the bottom of our 3d printer emissions testing page to see the reports, or download them below:


Comments on new working paper ‘Do Energy Efficiency Investments Deliver? Evidence from the Weatherization Assistance Program’

Economists Meredith Fowlie and Catherine Wolfram at the University of California, Berkeley and Michael Greenstone at the Energy Policy Institute at the University of Chicago (EPIC) have published a new working paper on the economics of residential energy efficiency retrofits, in which they present some discouraging results:

I first read the short research summary from EPIC, “Costs of Residential Energy Efficiency Investments are Twice their Benefits: Implications for Policy,” and found some surprising conclusions that led me to dig deeper into the actual working paper, “Do Energy Efficiency Investments Deliver? Evidence from the Weatherization Assistance Program.”

Continue Reading →

Gave a short presentation to ASCE Illinois today

I gave a presentation earlier today to the 3rd Annual Sustainability Workshop hosted by the Illinois chapter of the American Society of Civil Engineers (ASCE). Five different panelists briefly presented their perspectives on sustainability in their respective fields, then we had a short panel discussion. It was fun to talk energy, economic, and environmental sustainability of our country’s infrastructure with these folks and I was happy to be part of it!

I spoke about energy, environmental, and cost impacts of buildings in the US and some of the challenges to achieving widespread reductions in each area. You can download my short presentation here.

Literature round-up: Improving assessments of indoor exposures to outdoor air pollution

Much of my work over the last few years has been focused on improving methods to assess our indoor exposures to outdoor airborne pollutants. This is driven in part by the fact that we spend so much time indoors (and so much time at home) and that different pollutants can infiltrate indoors in different ways; at the end of the day, much of our exposure to outdoor air pollution actually ends up occurring indoors. Another motivator is that there are wide variations in some of the fundamental drivers of indoor proportions of outdoor pollutants, particularly in homes, that I don’t think have been captured very well to date. Air exchange rates are certainly higher in leakier buildings and we understand how to model these fairly well (although not as well as you might think), but good data on envelope penetration factors in a large number of homes are limited. We also don’t know a ton about HVAC filtration, system runtimes, indoor deposition rates, and the least academic yet probably most difficult of all to assess: window opening behaviors.

With all that said, we’re working on a handful of projects (and have a few related proposals under review) that would work to improve our knowledge of some of these drivers (as well as our predictive ability). Also, I noticed a couple of papers out in the Journal of Exposure Science and Environmental Epidemiology that continue to move this kind of work forward. I thought I’d share them here.

1. Breen et al. (2013) published a review of the models that can be used to estimate air exchange rates in buildings. They review driving forces (e.g., I/O temperature differences, wind speed, and mechanical ventilation) in conjunction with the leaks/openings through which driving forces can force airflow. Then they review a handful of models with varying levels of details and input parameter needs for estimating air exchange rates in buildings. They finish up with a list of advantages and disadvantages of each method and describe ongoing research needs. This is a very helpful paper for understanding how airflow can be assessed and ultimately used to impact exposure assessment for epidemiology.

2. Hodas et al. (2013) published a study where they used existing epidemiological data for myocardial infarction (heart attack) associations with elevated outdoor particulate matter (PM2.5). However, instead of using only outdoor concentrations, they accounted for variations in indoor proportions of PM2.5 across a variety of homes (related to the paper above!), as well as for time spent at home, in order to explore whether or not accounting for indoor exposures altered the outcome (accounting for differences in air exchange rates only). Interestingly, they didn’t observe differences in the exposure-response outcomes for those with different air exchange rates in their homes, but primarily because they used data from a “case-crossover” study whereby occupants are their own controls. That is to say that they track the same occupants in time with different exposures; therefore, it makes sense that indoor proportions of outdoor PM2.5 may be always consistent within a particular group, so building factors are in a sense controlled for in each “case”. They did however observe differences in the relative odds for heart attack for those in leakier homes, as occupants of leakier homes (with higher air exchange rates and thus higher indoor proportions of PM2.5 that would otherwise have not been accounted for using central site data). They conclude with “These ?ndings also illustrate that variability in factors that in?uence the fraction of ambient PM2.5 in indoor air (e.g., AER) can bias health effects estimates in study designs for which a spatiotemporal comparison of exposure effects across subjects is conducted.” Very much a motivator for the work we’re doing!

3. Baxter et al. (2013) actually preceded the study above by comparing the performance of a few different models for predicting indoor proportions of outdoor PM2.5 in homes — the models largely vary by the way they account for different air exchange rates in homes (related back to the first study here). Again, results suggest that differences in residential exposures may be important for epidemiology studies, and importantly, not captured by outdoor monitors alone. These kinds of studies again motivate me to get out there and develop better ways of assessing some of the inputs to these models (which are not always captured well) with our knowledge of building science to ultimately inform epidemiology studies and improve our decision making for regulatory purposes.


Summary of my first year (er, 9 months) as an assistant professor at IIT

So here we are, May 31st 2013. I’ve officially made it through my first two semesters as an assistant professor at IIT and I think it’s time to reflect briefly on the last year or so and assess how things went and where things are going! So here’s a quick rundown of what has gone on in and around IIT with me in the new job, lab, and students.

I signed on at IIT in late March 2012 and officially started as an adjunct assistant professor in May 2012. I pursued adjunct status immediately so I could send out a few research proposals last summer before starting officially in August (so I’ve only been with IIT about 9 months, depending on how you count it). Since then, the lab I share with Prof Paul Anderson has been completely renovated and is in the best shape it’s been in for a very long time (our building was built in the mid 1940s). We split the lab space into student offices on one side and lab facilities on the other side, although my use of “lab” is very different from others that work in much more controlled environments — we don’t even have a fume hood. We use the lab space primarily as a staging ground to assemble hardware and sampling rigs for field campaigns. Anyway, it’s a very functional space now and it warms my heart to see students working away on various projects in there! You can follow some of the previous progress here, although I still need to update with some newer photos.

Now, let me quickly summarize some of the past year’s worth of research, teaching, and student achievements.


In the past year I’ve spent a lot of time working on a few new research projects. In the first year I’ve received funding for the following three projects:

  • Hospital Microbiome Project. Our lab is funded by the Alfred P. Sloan Foundation to measure a variety of building science parameters in a brand new hospital that opened at the University of Chicago. The utmost goal of the project is to characterize microbial communities as they develop and evolve in a new hospital right before it opens and as it is occupied with staff and patients for a year. We are working to characterize building operation and environmental parameters so we can hopefully learn something about how buildings influence microbial communities in a critical environment like this hospital. Jack Gilbert at UC/Argonne Lab is the PI on the project and we are a subcontract. We recently attended a meeting hosted by the University of Colorado and funded by the Sloan Foundation to present some of our initial measurements and findings. Our work has been well received in the microbial ecology community and we look forward to continuing to work on this for the next 9-12 months! Also stay tuned to the Microbiology of the Built Environment Network (microBEnet) blog for continued updates in this field.
  • Life cycle costs of duct designs in residences. Switching gears a bit to energy use in the built environment, we’ve also been working on a project funded by the Air-conditioning, Heating and Refrigeration Institute (AHRI) to model the energy impacts and overall life cycle costs of a range of low, medium, and high pressure HVAC duct designs in homes in Chicago, IL and Austin, TX. The idea here is that ductwork designs can introduce excess system pressures that an air handling unit fan has to overcome in order to distribute conditioned air to the space. Those pressures may have some energy impacts but designing lower pressure duct systems also costs money! We’re just wrapping up the draft final report on this work and will update its project page as we complete it. We’ll also prepare this work for publication in the near future (i.e., this summer).
  • HVAC filtration for reducing risks of airborne infection in indoor environments. In another interesting project funded by the National Air Filtration Association (NAFA) Foundation, we’ve been working to model the potential impacts that HVAC filters might have on controlling infectious aerosols and thus reducing risks of spreading airborne infectious diseases in indoor environments. I’ve personally learned a lot through this work, which involved digging deep into the literature on infectious disease epidemiology, biology, and the physical processes that govern the transmission of diseases like influenza, tuberculosis, and rhinovirus (the common cold). One of our doctoral students, Parham Azimi, and I just prepared a manuscript on this work for submission. Also, the full report is available online. Parham continues to work to compare the risk model we used with other methods of modeling infectious aerosol dynamics and risk in indoor environments.

Aside from our funded projects, I’ve also worked closely with students on a range other unfunded projects. One of our masters students, Zeineb El Orch, has been working on a really nice Monte Carlo model of the infiltration of outdoor particulate matter inside homes in the US. She’s working on her thesis and a couple of publications from this work as we speak. In our first short field campaign, a group of us went out to a local 3D printer shop and performed measurements of ultrafine particle emissions from desktop 3D printers. We’ve already submitted this work for publication and it’s under review now — it turns out some of these printers emit a lot of particles in the nanoparticle/ultrafine size range and they appear to differ quite drastically depending on the plastic feedstock used and the operating temperature. We also published a paper with a Ph.D. student in the College of Architecture (now Dr. Irina Susorova) on the development and validation of a model she built for evaluating the thermal performance of ivy on exterior walls of buildings. We’ve also been acquiring equipment and supplies and gearing up for field work for measuring indoor and outdoor size-resolved particle and gas-phase pollutant inftilration measurements. 


In the past year I’ve also spent a lot of time preparing lectures and teaching my first two classes full time, as is tradition for new faculty. In Fall 2012 I taught my first course: CAE 463/524 Building Enclosure Design (mixed undergrad and grad). Then in Spring 2013 I taught my second course: ENVE 576 Indoor Air Pollution (grad). I enjoyed both and learned a lot from both — about the material itself as well as about my own teaching styles (whatever those are) and student learning. I post all of my lecture notes online so you can get a sense of what we’re covering. I intend to keep teaching Indoor Air Pollution once per year for, oh, the rest of my life. I really enjoyed it. I’ll probably keep teaching Building Enclosure Design as well — it’s all about heat, air and moisture transport through building envelopes (roofs, walls, windows, floors) and designing them in a way to control those flows. My only real complaint with this course is that I felt like I wasn’t always teaching to the right discipline. We really need to increase the number of architecture students who take this class because in the way that we do things, architects often make these envelope detail decisions and by the time the drawings get to engineers it may be too late to do anything! We’re working on this in part by working with the PhD program at the College of Architecture, which leads us into a summary of collaborative efforts….


In the past year I’ve also spent a lot of time working on building my network of collaborators both within IIT, around Chicago, and elsewhere. I’m thankful to have begun working with some students in the College of Architecture at IIT, particularly in their PhD program. We’ve been discussing ways to grow our programs together. I’ve also worked a bit with a couple of other faculty members within IIT, from mechanical engineering to transportation to applied math. I hope these collaborations continue to grow. We’ve also met a few times with CNT Energy — a nonprofit “think and do tank” in Chicago whose model I really respect. Same with some faculty members at UIC (e.g., Rachael Jones in public health) and with a practitioner colleague, Ian Cull at Indoor Sciences. I’ve been PI on a couple of proposals that includes these folks so hopefully some will hit and we’ll get to work together for real.

We’re also keeping up good relations with my PhD adviser Jeff Siegel, now at the University of Toronto, and with Michael Waring at Drexel University, among others. I’m always open to collaboration and I think some of our research work in a wide variety of areas will continue to push us here.


Also in the past year, we’ve grown the number of students working in the lab, from zero students in the beginning to as many as about 10, depending on how you count them all! Our current students are a mix of bachelors, masters, and doctoral candidates from a range of disciplines including civil engineering, architectural engineering, environmental engineering, mechanical engineering, chemical engineering, and architecture. Congrats are already in store for a couple of graduates as well: Irina Susorova (PhD Architecture) and Ben Wachholz (BS Civil) both graduated in May! Also congrats are in store for Tiffanie Ramos (MS Environmental) who won a prestigious Starr-Fieldhouse Fellowship within IIT for her work with the Hospital Microbiome Project team at Argonne Lab.

I’ve also been somewhat active in working with some great student groups. One group I’m working with is building a greenhouse facility for a school on Chicago’s south side. I’ve also spoken with members of the Engineers Without Borders and ASHRAE chapters at IIT to some extent. We’ve also participated in the College of Engineering’s undergraduate research program where some undergrad students have gotten to cut their teeth on research and actually contribute to some of our research work and lab development. Overall, working with my students is really a joy — but I’ve also learned that mentoring and management is a skill that takes time to hone. Hopefully for their sake I’m getting there! I’ve also realized that watching students graduate this spring was even more enjoyable than I thought! … it was great to see everyone graduate and move on to bigger and better things!

Plans for the next year

As I move forward into my second year and beyond, I’m looking forward to continuing to pursue a wide range of research interests. We are going to work to crank out 4-5 publications this summer now that I have some time to continue to generate content here and summarize the work we’ve been doing over the last 9 months. I’ll also be adding a new course to my teaching portfolio this fall: CAE 331/513 Building Science, which I’m looking forward to. I’ll also be submitting an application to develop an M.S. program in architectural engineering (currently we only have a professional masters, which is a non-thesis, non-research degree). 

I also need/want to blog more on this site (which was part of the motivation for this post!). I started using an offline blog editor (MarsEdit) that looks like it will encourage me to post more. We’ll see. I will try to post at least weekly and will continue to generate my “literature round-ups” that cover some interesting journal articles that I come across on a day-to-day basis. I’m also pushing to put some of the things we learn in the lab onto a group wiki for all students to use — I have dreams of this increasing efficiency and productivity for all as we move forward!

Until next time, take a look at our projects and publications and follow us (er, me) on Twitter @built_envi. And never hesitate to reach out and/or stop by the lab to chat!

Dr. Stephens gives EWB-IIT lecture on indoor air pollution in developing countries

It’s a little late for this as a news item, but I wanted to post slides here. I gave a lecture to the IIT chapter of Engineers Without Borders USA (EWB-USA) today at their general body meeting. I covered some of the basics of the huge global health burden attributed to indoor and household air pollution primarily in developing regions of the world. Some EWB members had recently been in Nicaragua on a separate bridge-building project and happened to notice women in the community cooking on inefficient cook stoves. They also noticed their interior walls were covered with soot — a pretty solid lead-in to this presentation. Hopefully they will be able to work on some cleaner stove intervention projects in the future!

My slides are available here.

Group member wins internal IIT research fellowship!

The good news was finally announced on IIT’s website but we’ve known for a couple of months: research group member Tiffanie Ramos (MS ENVE expected Fall ’13) earned the prestigious Starr-Fieldhouse fellowship at IIT! The fellowship supports students who are working on research jointly between IIT and one of a few surrounding research institutions, including Argonne National Lab, Fermi Lab, or IIT’s private Research Institute. Tiffanie was awarded the fellowship for her proposal to support the ongoing Hospital Microbiome Project, which we are working on with PI Jack Gilbert at the University of Chicago and Argonne Lab (read more here). She will be assisting in building science measurements to support air sampling and better characterize environmental conditions, HVAC operation, and human occupancy in patient rooms of a new hospital (meanwhile, the ANL team is sampling ~13,000 surfaces over the course of a year to explore how microbial communities change over time once a new hospital is occupied). Congrats to Tiffanie!

Now if you’ll excuse us, we have to go setup equipment at the hospital!

Added two new project descriptions to the site

I’ve added descriptions of two new projects that we’ve begun this fall! They are described on the Projects page. These two projects include:

Stay tuned as we conduct the research for these projects and post our results!