Monthly Archives: November 2008

Week 13: Carbon calculators

Since I’m storing all my literature review information my LaTeX document, there are a fair number of URLs listed. I figured there must be a way to get LaTeX to generate PDFs with clickable links, and sure enough there is. The hyperref package has lots of nifty stuff, including making URLs clickable and making internal document references also clickable. Since I’m using this LaTeX Makefile, which produces output via LaTeX -> DVI -> PS -> PDF, I also need the breakurl package.

Here’s the code I’m using, which should be the last stuff before the \begin{document} :

%% Make URLs clickable
\usepackage[colorlinks, bookmarks=true]{hyperref}

%% Since I'm using the LaTeX Makefile that uses dvips, I need this
%% package to make URLs word wrap nicely
\usepackage{breakurl}

Preliminary results from the carbon calculators are interesting. Using my own personal data, Carbon Counter believes my emissions from air travel this year total 20.58 metric tons of CO2, while Dopplr‘s estimate for my air travel plus all the driving during my road trip totals 5.87 metric tons of CO2, a very different number. Carbon Counter sells carbon offsets (their calculator seems designed mostly as an input to tell you how many carbon offsets you should buy), so it would be in their financial interest to skew results higher. However, it’s too early to tell why the calculations are different. Carbon Counter has a nice page explaining how they calculate the values, and I haven’t investigated how AMEE (the service used by Dopplr) does the calculations.

No word from Dopplr on whether they would add a method to their API to extract the carbon data from trips. If they are unable or unwilling to do so, I could use the API to extract the travel information and send it to AMEE myself.

Other than continuing my reading, the big task for this shortened week is to crank out a first draft of my tech report.

Planned items from last week:

  • Read 2 new papers from literature review list
    • [DONE] Leveraging Social Networks To Motivate Individuals to Reduce their Ecological Footprints
    • [DONE] Do increases in energy efficiency improve environmental quality and sustainability?
  • Write up notes from papers read during PET paper lit review
    • [DONE] Challenging Comfort & Cleanliness Norms Through Interactive In‐Home Feedback Systems
  • [STARTED] Run through carbon calculators with my data
  • [DONE] Fix LaTeX table
  • [DONE] Follow up on Dopplr API request

Other accomplishments this week:

  • Figured out how to use ssh-agent on Leopard, so my script to sync my proposal and tech report doesn’t prompt me for my password more than once per session
  • Figured out how to make hyperlinks in LaTeX
  • Dug through old email messages for additional things to read, added to queue

Hours worked: 17 (target: 15 hr)

Plans for coming week:

  • Write first draft of tech report
  • Read 2 new papers from literature review list
    • Carbon neutral – sense and sensibility
    • How Much Energy Are We Using? Potential of Residential Energy Demand Feedback Devices
  • Write up notes from papers read during PET paper lit review
    • Promoting Environmentally Sustainable Behaviors Using Social Marketing In Emerging Persuasive Technologies
  • Finish running through carbon metric calculators

Pointers to work products:

Cool links:

  • Nothing of note

ssh-agent and Mac OS X 10.5 Leopard

I’ve been using ssh-agent since I started using Mac OS X (10.1), it being the best combination of security and convenience when using ssh and scp/sftp. During that time I have been using some modifications to my .tcshrc file to start ssh-agent when I log in and set up the ssh environment variables properly.

However, Mac OS X 10.5 aka Leopard now has nice ssh-agent support built-in that my startup file was overriding. I wanted to understand more about the new ssh-agent support in Leopard before I went removing my ssh-agent code. Sure enough, there are a variety of blog posts that explain the situation. This post at Ormset i Noreg explains the situation, and discusses complications if you have been using the open source SSHKeyChain GUI application to manage your ssh-agent under Tiger.

Dave Dribin goes into more depth on exactly what is going on with the keychain. He also has a followup post on how to keep your passphrase stored in a separate keychain to make it less vulnerable that I plan to follow. He also has written some code to support one of SSHKeyChain’s features: removing keys from the agent when the computer is put to sleep.

It would be nice if SSHKeychain was rewritten to be more of a support service for the Leopard SSH stuff, but based on some discussion on the mailing list it appears that the author has been looking for a new maintainer for a few months.

Week 12: Dopplr data

I spent some time working with Dopplr this week. It now has email import features of the type that make TripIt so popular. I haven’t tried email input yet since I’m not traveling. There is a Dopplr API, but it doesn’t appear that the carbon footprint information is available via the API. I have submitted a request to add carbon querying to the API, and someone from AMEE has replied that Dopplr would need to expose the AMEE ID and then the data could be directly downloaded from AMEE. AMEE is looking like it might be a core part of PET, since they do carbon footprint calculations for all sorts of input data.

Someone created something called Offsettr that uses the Dopplr API to calcuate carbon footprint information for your entire Dopplr “fellow traveler” network from the trip distance information, but it assumes all travel is by plane.

I already had input all 3 of my trips this year into TripIt, so I was looking for a way to import them into Dopplr. The officially suggested method for doing this is to have Dopplr subscribe to TripIt’s iCal feed (also described in slightly more detail elsewhere), so I tried that. It worked OK for trips where you fly into one place and then fly out, but it only picked up the first leg of my multi-hop road trip in the Pacific Northwest. After considerable manual editing (and some nasty issues with the Dopplr web interface), I got that all entered properly. I then checked out the carbon footprint it calculated for me: 5,871 kg CO2! Yipes! Here’s the graphic:

My carbon footprint calculated by AMEE via Dopplr

My carbon footprint calculated by AMEE via Dopplr

Other ideas I had this week. The Darby paper on the effectiveness of feedback on energy conservation got me thinking about ambient displays. How should one display information about usage (electricity, etc) that is informative, but uses very little power for the display? Large computer or TV displays are the canonical choice, but that uses a fair amount of electricity. Perhaps something mechanical that moves, like a gauge or clock? Electronic ink?

Getting electricity usage for a particular appliance is easy (Kill-A-Watt), and getting usage for a whole house is also straightforward (The Energy Detective). But ideally you’d like information about total information usage plus usage data for every device plugged in. Maybe a total energy meter plus something like what’s described in At the Flick of a Switch. The Device-Level Power Consumption Monitoring paper describes a wireless system for aggregating data from plug loads throughout an area, but that will require a bunch of the meters, one for each device you want to monitor.

It would be great if HECO had live energy graphs like these ones that LBNL created during the California power crisis that showed the current usage, forecast usage, and capacity. It could also break down the capacity based on different types of generation: oil, coal, wind, etc.

Planned items from last week:

  • Read 2 new papers from literature review list
    • [DONE] New Ways to Promote Proenvironmental Behavior: Expanding and Evaluating Motives for Environmentally Responsible Behavior
    • [DONE] The effectiveness of feedback on energy consumption
  • Write up notes from papers read during PET paper lit review
    • [DONE] Device-Level Power Consumption Monitoring
  • [DONE] Put carbon footprint calculators in table form in LaTeX
    • The table exists, but the URLs are too long so the table goes off the page. Gotta fix that.
  • [NOT DONE] Run through carbon footprint calculators with personal data
  • [DONE] Check out Dopplr’s carbon tracking for travel

Other accomplishments this week:

  • Requested Dopplr add API method for retrieving carbon footprint

Hours worked: 15 (target: 15 hr)

Plans for coming week:

  • Read 2 new papers from literature review list
    • Leveraging Social Networks To Motivate Individuals to Reduce their Ecological Footprints
    • Do increases in energy efficiency improve environmental quality and sustainability?
  • Write up notes from papers read during PET paper lit review
    • Challenging Comfort & Cleanliness Norms Through Interactive In‐Home Feedback Systems
  • Run through carbon calculators with my data
  • Fix LaTeX table
  • Follow up on Dopplr API request

Pointers to work products:

Cool links:

  • Get Satisfaction is a web site that allows users to ask questions and post ideas about companies. Companies can register employees who can post official responses, and users can help other users. It’s an outsourced social support service, basically.

Eating some meat & dairy might reduce environmental footprint

Just saw this article about how eating some meat might reduce one’s overall environmental footprint. The study found that while a vegetarian diet used the least amount of land area, it required all the land area to be high-quality agricultural land. Adding a small amount of meat and dairy to that diet uses more land, but some of that is pasture land, which is more abundant in the New York state area and necessary for crop rotation anyway.

Note that the land required for the typical American diet of lots of meat is much greater than either scenario above. So no surprises there.

Week 11: defining carbon footprint

I’ve been talking about “carbon footprint” as if it were a well-defined concept, but it turns out that it can be hard to nail down. I read a report that delves into how to define carbon footprint: do we include methane? what about NO2? what about emissions that don’t come from burning fossil fuels? There’s also the question of what units to represent carbon footprint in (some have suggested that it be measured in the amount of area of forest required to sequester the emitted CO2). In the end the authors come up with this definition:

The carbon footprint is a measure of the exclusive total amount of carbon dioxide emissions that is directly and indirectly caused by an activity or is accumulated over the life stages of a product.

They also discuss the methods one can use to compute carbon footprints. Perhaps unsurprisingly, calculating something’s carbon footprint is difficult. For a product or process, a bottom-up approach is suggested, but this doesn’t capture all of the indirect emissions from the supply chain. The environmental input-output method can take into account all the higher-order impacts, but works best at the more macro level of a government or industry sector.

Per Philip’s suggestion, I created a separate CSDL tech report that I will use to take a snapshot of the literature review from my proposal at the end of the semester. This snapshot will be submitted as my final report for ICS 699 this semester, and also will make up a major portion of my research portfolio (to be submitted in January 2009). To accomplish this I removed the content of the literature review chapter from my proposal body LaTeX file and put it in its own file. I left the \chapter{} in the proposal body since I don’t want that included in the tech report. However, the \include{} command does not allow chaining and the body of the proposal is already included into the main document. This helpful page (why are most LaTeX FAQs and tutorials written by graduate students working on their dissertation?) explained that \input is the other way to insert content from another file, and it allows chaining and doesn’t create a new page, which is exactly what I wanted.

Using symbolic links, the tech report includes the content of the literature review from the proposal, and also the BibTeX file. When I make the snapshot at the end of the semester, I’ll “break” the symlinks and replace them with copies of the linked file to ensure that the tech report doesn’t keep changing as the proposal changes.

Planned items from last week:

  • Read 3 new papers from literature review list
    • [DONE] Learning Transportation Mode from Raw GPS Data for Geographic Applications on the Web
    • [NOT DONE] New Ways to Promote Proenvironmental Behavior: Expanding and Evaluating Motives for Environmentally Responsible Behavior
    • [DONE] A Definition of ‘Carbon Footprint’
  • [NOT DONE] Check out Dopplr’s carbon tracking for travel
  • Write up notes from papers read during PET paper lit review
    • [DONE] Motivating Environmentally Sustainable Behavior Changes with a Virtual Polar Bear
  • [DONE but not in table form] Start list of carbon footprint calculators, shown in table form in proposal document

Other accomplishments this week:

  • Created a separate CSDL tech report 08-03 to be submitted as my final report for ICS 699 this semester, and to include in my research portfolio. It uses the same underlying content as the literature review chapter from my proposal via symlinks. Isn’t LaTeX great?
  • Wrote quick script to backup my proposal (and tech report) directory to dasha via rsync

Hours worked: 15 (target: 15 hr)

Plans for coming week:

  • Read 2 new papers from literature review list
    • New Ways to Promote Proenvironmental Behavior: Expanding and Evaluating Motives for Environmentally Responsible Behavior
    • The effectiveness of feedback on energy consumption
  • Write up notes from papers read during PET paper lit review
    • Device-Level Power Consumption Monitoring
  • Put carbon footprint calculators in table form in LaTeX
  • Run through carbon footprint calculators with personal data
  • Check out Dopplr’s carbon tracking for travel

Pointers to work products:

Cool links:

  • Doodle is a super quick way to do informal surveys. The primary thing I use it for is coordinating a meeting time between a group of people. You select the set of days and time slots during those days, send the URL out to the people involved, and Doodle will compute which day and time works best. I especially like that you don’t have to create an account to make or participate in a poll, which makes it viable for scheduling with busy people.
  • DropSend is one of the many services designed to let you email large files to people. You send the email through their web interface, the large file (up to 1 GB) is uploaded to their server, and the recipient gets an email with the (temporary) URL to the file. Drop.io is another data storage service, which allows data input via a variety of methods but allows file storage without creating an account. The downside is that their free service only allows 100 MB, which was too small for my immediate needs.

Week 10: thoughts on air travel

This week was more reading from my ever growing reference list. I’ve now organized my list of possible things to read by sections since it was getting pretty cluttered. It comes at the end of my Related Work chapter. I’m finding that a LaTeX document isn’t the greatest place to collect links, but it’s workable.

I came across this somewhat depressing article on how green consumers tend to be major carbon emitters. People who were most environmentally conscious (recycling, reducing car usage, etc) were also the one’s most likely to take long airplane flights, which wipes out any savings made in the home. I have to track down the actual paper, but it sounds like the environmentally conscious folks were the wealthiest, thus giving them the ability to take long flights. I hope that having a unified interface such as the one proposed in PET would help people to see the whole picture of their environmental impact.

If airplane flights really are the dominating factor in carbon emissions for most middle-class people, then that has a particularly nasty impact for Hawaii residents, since we live 2000 miles away from almost everything. Other than flights to neighbor islands, all flights are long flights and there is no practical alternative to flying. However, if flights are the biggest problem, then that could provide renewed focus on ways to avoid air travel. While it’s hard to imagine a realistic IT-based alternative for vacations, advances in telepresence could conceiveably make flying away for business meetings unnecessary.

My love affair with BibDesk continues. This past week I discovered the web import functionality. BibDesk provides an embedded WebKit browser that you can use to search a website like the ACM Portal for papers. When you find one that you are interested in, BibDesk can import the bibliographic data with one click. Nice.

Per discussion with Philip, need to read 3 papers a week to stay on target for literature review. At the end of the semester, I’m going to take a snapshot of the literature review chapter and turn that into a separate tech report that I’ll turn in as the document for ICS 699 and as a portion of my research portfolio.

Planned items from last week:

  • Read 4 new papers from literature review list
    • [NOT DONE] Learning Transportation Mode from Raw GPS Data for Geographic Applications on the Web
    • [DONE] Energy efficiency: rebounding to a sound analytical perspective
    • [NOT DONE] New Ways to Promote Proenvironmental Behavior: Expanding and Evaluating Motives for Environmentally Responsible Behavior
    • [NOT DONE] A Definition of ‘Carbon Footprint’
  • [NOT DONE] Add more sensors to list
  • Write up notes from papers read during PET paper lit review
    • [DONE] EcoIsland: A System For Persuading Users To Reduce CO2 Emissions
    • [DONE] Taking the Guesswork out of Environmentally Sustainable Lifestyles

Other accomplishments this week:

  • Emailed author of Understanding Mobility Based On GPS Data

Hours worked: 12 (target: 15 hr)

Plans for coming week:

  • Read 3 new papers from literature review list
    • Learning Transportation Mode from Raw GPS Data for Geographic Applications on the Web
    • New Ways to Promote Proenvironmental Behavior: Expanding and Evaluating Motives for Environmentally Responsible Behavior
    • A Definition of ‘Carbon Footprint’
  • Check out Dopplr’s carbon tracking for travel
  • Write up notes from papers read during PET paper lit review
    • Motivating Environmentally Sustainable Behavior Changes with a Virtual Polar Bear
  • Start list of carbon footprint calculators, shown in table form in proposal document

Pointers to work products:

Cool links:

  • Checks are now (apparently) the most dangerous way to give someone money. Every check has both the bank routing number and your account number, and with that information a criminal can pull money out of your bank account with far fewer protections than if they were stealing your credit card number.