Spaced Out :: 05 2014

Came across this nice piece over at The Atlantic on taking lecture notes…. the short takeaway is that writing lecture notes helps you remember them more easily. However lets dig in to the report a little further….

They conducted a trial: (1) Watch lecture and take notes, (2) Complete irrelevant mental tasks for 30 minutes and (3) Take a quiz on the material. Step (2) is important to allow the “forgetting curve” to kick in (see here). The key finding, as reported, is this:

“In this group, longhand-notetakers outperformed laptop-notetakers on the quiz. Analysis of student notes showed that laptop-notetakers tended to transcribe a lot of the speaker’s words verbatim. Mueller and Oppenheimer suspected that this was because those who typed notes were inclined to transcribe lectures, rather than process them. This makes sense: If you can type quickly enough, word-for-word transcription is possible, whereas writing by hand usually rules out capturing every word.”

Being able to type quickly means that you become mentally lazy - the key point which was NOT drawn out is this…. deep learning takes place when you assimilate new ideas, classify them, interpret them, interject them with your own thoughts, rework them in to your mindset and then see where you started. Surficial learning occurs when you simply try to “record” what is being said - this could be by recording the audio and then just listening, or it could be writing it all down, verbatim, word-for-word. This is knowledge, not learning. If all you want to do is this, buy a book, read it. Don’t waste your time going to a lecture!! If however you really do want to learn, treat the lecture as an interactive medium - the lecturer shouldn’t be there to simply read out a set of notes to you, but rather introduce you to new ideas, concepts or techniques, get you to practice them, think about them, assimilate them…. reframe what you hear within your own understanding or context, ask questions in order for it to make sense, be classified, within your mindset. In the genuine sense of the word, your are turning your ignorance in to learning, but it is only with that active participation that this can happen.

So it’s not about laptops or writing, but rather our approaches to learning…unfortunately it appears many students can’t make that distinction, even when they are told about it.

Nice video over at the BBC on the new Parrot HD drone - yes, wifi controlled, fisheye lens, 2km range and the virtual reality Occulus Rift headset to give a bird’s eye view!

cube sats… obviously! Planet Labs Flock-1 has been in the news quite a bit since it successfully launched its initial 2 cube sats from the International Space Station. These are remarkable little satellites (see the image sequence on the link above at the bottom of the page) that are designed to be small, light, very low cost and replaceable, allowing the provision of relatively high resolution imagery (~4m) of the entire globe at any one time using off-the-shelf components. Each unit is called a Dove and is about the same size as one! Weighing in at only 5kg. But its the IT side which may well be the bigger task - downloading, stitching and distributing a 10 terapixel image of a regular basis is no trivial task!

UPDATE: and hot on the heels, covered over at the BBC and this

Another piece of news that slipped past the radar (O dear O dear….)…. JAXA is intending to produce another DEM, this time from ALOS PRISM data. ALOS (or Daichi) was launched in 2006 and then decommissioned (due to failure) in 2011 (see Wikipedia entry). Three instruments are on board ALOS: PRISM, PALSAR and AVNIR-2. PALSAR collects radar imagery, whilst AVNIR-2 is a multispectral radiometer with 10m spatial resolution. It is PRISM that is of particular interest as its design to collect stereo data for DEM production - it has three 2.5m resolution sensors (fore, aft and nadir) which collect stereo data.

The global mapping project aims to generate a 5m DEM of the world, with a secondary coarser resolution 30m DEM which will be free to use. There has been quite a lot written on using raw PRISM data (for example this paper; try Google Scholar) although it seems common to generate 10m pixels with an accuracy of 3-5m. Processing of the global DEM will take 2 years and my expectations are high that this will significantly boost detail of topographic features of the Earth’s surface.

This is one of those things you just think should be easy (one line of code in R!) and turns out to be not quite so simple!! The solution was culled from this page and this page. The “work around” is to utilise the secondary y-axis in Excel to show two values that are the same, scaled from 0 to 1, then hide the axis labels. The quick steps to this are:

1. Create a dummy series such as
0 865
1 865

2. Goto Chart Source Data: add this as a new series (x: 865 points, y: 0 1 points)

3. Goto to Format Data Series (double click on the line itself): goto the Axis tab and click Secondary axis. This will assign it to a secondary *y-axis”

4. Goto Chart Options: click the Axes tab and, voila, both secondary x and y axis options are available. For the secondary axis tick x-axis

5. Scale Axes: for the secondary x axis (double click the axis itself) goto the Scale tab and make sure it goes from a minimum of 0 to a maximum of 1. For the secondary y axis goto the Scale tab and make sure it matches the scale of the primary y axis.

6. Format Axes: for *both* secondary axes goto the Patterns tab and make sure that all tick mark types/labels are set to none in order to turn the labeling off. On the Scale tab untick “Value (Y) axis crosses at maximum category” and “Value (Y) axis crosses between categories”

That should do it - tested in my latest version of Microsoft Excel - yes, 2003!!