convore.json/groups/python-scientific-computing/data-array-in-numpy/messages.json

[{"user_id": 11207, "stars": [], "topic_id": 34736, "date_created": 1305118779.5806341, "message": "We are hosting a summit this weekend on moving the ideas of the data array (larry, pandas, etc.) into core NumPy and improving Python for statistical data analysis this weekend (May 13 and 14) at the Enthought offices in Austin.  If you are in Austin, and interested in contributing to the discussion / sprints, please come.", "group_id": 6727, "id": 1006931}, {"user_id": 11207, "stars": [], "topic_id": 34736, "date_created": 1305118790.2585869, "message": "Here is a link to some ideas about format / agenda:  https://docs.google.com/document/d/1f-FknatNu-3iguZd3Nl4AAhrDTb-PoOh4mVNziq0rss/edit?hl=en&authkey=CL2YwMEO", "group_id": 6727, "id": 1006933}, {"user_id": 11207, "stars": [], "topic_id": 34736, "date_created": 1305118822.149441, "message": "Feel free to comment and/or send feedback on what we should be discussing at the event --- if you can't attend, we will use Convore to chat about the ideas as well.", "group_id": 6727, "id": 1006938}, {"user_id": 14878, "stars": [], "topic_id": 34736, "date_created": 1305125712.5719349, "message": "I was not aware of the larray and panda and other proposal. I found these 2 pages http://projects.scipy.org/numpy/wiki/NdarrayWithNamedAxes and http://scipy.org/StatisticalDataStructures Are they relevant to the discussion as well?", "group_id": 6727, "id": 1008442}, {"user_id": 11207, "stars": [], "topic_id": 34736, "date_created": 1305198992.0149491, "message": "Yes.  Those are absolutely relevant links.   Thanks for posting them.", "group_id": 6727, "id": 1023755}, {"user_id": 10411, "stars": [], "topic_id": 34736, "date_created": 1305217081.2514589, "message": "Also, could someone who gets to go to this please wrangle up a few people for an inSCIght recap episode, please?  I would be ever grateful...", "group_id": 6727, "id": 1025987}, {"user_id": 8649, "stars": [], "topic_id": 34736, "date_created": 1305239616.9848011, "message": "Any discussion along the lines of labelled/tagged data, Numpy record and stats data arrays, etc. would not be complete without doing some background reading on Essbase: http://en.wikipedia.org/wiki/Essbase", "group_id": 6727, "id": 1030955}, {"user_id": 8649, "stars": [], "topic_id": 34736, "date_created": 1305239650.1043091, "message": "As long as we are thinking about trying to solve these problems, it's worth looking at the thought leaders and current front-runner commercial offerings in the space of hypercubes, MDX, etc.", "group_id": 6727, "id": 1030966}, {"user_id": 33184, "stars": [], "topic_id": 34736, "date_created": 1305301138.8038061, "message": "@jonathanrocher those pages are highly relevant to the discussion, the latter one was my doing", "group_id": 6727, "id": 1043164}, {"user_id": 21711, "stars": [], "topic_id": 34736, "date_created": 1305303191.9126389, "message": "for kdb", "group_id": 6727, "id": 1043567}, {"user_id": 21711, "stars": [], "topic_id": 34736, "date_created": 1305303191.730577, "message": "http://cosy.com/language/k-lang.htm, says 100k per developer/year", "group_id": 6727, "id": 1043566}, {"user_id": 11207, "stars": [], "topic_id": 34736, "date_created": 1305309009.409919, "message": "https://docs.google.com/a/enthought.com/?ndplr=1&pli=1#folders/folder.0.0B_u3gF4pEw9INzBlMzE1NjYtMTY0NC00MmI2LTk0YzYtZjExZDdmMjgzYzk4", "group_id": 6727, "id": 1044635}, {"user_id": 11207, "stars": [], "topic_id": 34736, "date_created": 1305309064.2594471, "message": "This is is a folder were some notes are congregating.", "group_id": 6727, "id": 1044641}, {"user_id": 8649, "stars": [], "topic_id": 34736, "date_created": 1305316294.1681421, "message": "Python 3j, you heard it here first", "group_id": 6727, "id": 1046343}, {"user_id": 11207, "stars": [], "topic_id": 34736, "date_created": 1305317064.6341541, "message": "That was hilarious....", "group_id": 6727, "id": 1046532}, {"user_id": 11207, "stars": [], "topic_id": 34736, "date_created": 1305317082.6953599, "message": "We should definitely push for Python 3j", "group_id": 6727, "id": 1046536}, {"user_id": 33183, "stars": [], "topic_id": 34736, "date_created": 1305321889.371753, "message": "If you all go do something after/around that time please text me!  (520)591-5245", "group_id": 6727, "id": 1047593}, {"user_id": 8649, "stars": [], "topic_id": 34736, "date_created": 1305322058.394654, "message": "ok will do.  we'll almost certainly grab beers.  GL with your final!", "group_id": 6727, "id": 1047621}, {"user_id": 33183, "stars": [], "topic_id": 34736, "date_created": 1305321865.680861, "message": "Sorry all for ducking out.  Got to study for a final at 7-10ish", "group_id": 6727, "id": 1047588}, {"user_id": 33183, "stars": [], "topic_id": 34736, "date_created": 1305381426.6442111, "message": "Are things starting at 10 today per the notes?", "group_id": 6727, "id": 1052554}, {"user_id": 21711, "stars": [], "topic_id": 34736, "date_created": 1305387094.508626, "message": "yup", "group_id": 6727, "id": 1052924}, {"user_id": 21711, "stars": [], "topic_id": 34736, "date_created": 1305476570.310009, "message": "flight delayed, they're looking for parts...", "group_id": 6727, "id": 1060627}, {"user_id": 8649, "stars": [], "topic_id": 34736, "date_created": 1305477539.4618411, "message": "that's sometimes code for \"the pilot overslept\"", "group_id": 6727, "id": 1060732}, {"user_id": 8649, "stars": [], "topic_id": 34736, "date_created": 1305477613.10588, "message": "he'll be refreshed for the flight", "group_id": 6727, "id": 1060749}, {"user_id": 10454, "stars": [], "topic_id": 34736, "date_created": 1305477581.733469, "message": "Also not a good sign.", "group_id": 6727, "id": 1060741}, {"user_id": 10454, "stars": [], "topic_id": 34736, "date_created": 1305477446.3217471, "message": "Never a good sign.", "group_id": 6727, "id": 1060723}, {"user_id": 11207, "stars": [], "topic_id": 34736, "date_created": 1305574695.1431949, "message": "Here is my bullet-list of highlights about what went on in the conference: \n\t* acute recognition of the need to make the date-time support work in NumPy\n\n\t* discussions with Mark Wiebe and others about NumPy 2.0 changes including adding data-array features directly to indexing.  Other changes include: \t\n\t       * fancy-indexing producing views\n\t       * adding ability to do calculations with structured arrays\n\t       * indexing with the structure \"dimension\" \n\t       * cleaning up some of the other behaviors of NumPy\n\n\t* agreed upon the API for adding data-arrays to NumPy:   adding a single .axes attribute which is an object that allows attribute-access to each of the named axes as well as labeled indexing. \n\t    Example: \n\t\t      arr.axes(\"stocks\", \"metrics\")['goog':'ibm', ['open', 'high', 'close']]\n\n\t* saw the need to write a fast and robust text-file reading tool for NumPy and started discussions on it.   It should just work\n\n\t* some discussions on making sure we get an easy way to do quick, but powerful regression analysis based on data stored in a NumPy array.  \n\n\t* some discussions on interpolation (often indexing is a useful paradigm for getting data that is not really part of the data but only implied by the data if you assume some interpolation strategy) --- how would this work and should it be in NumPy. \n\n\t* some discussions on \"generated\" arrays including \"generated\" columns in an array", "group_id": 6727, "id": 1073943}, {"user_id": 11207, "stars": [], "topic_id": 34736, "date_created": 1305574721.544507, "message": "by conference, I mean the data array summit this weekend.", "group_id": 6727, "id": 1073950}, {"user_id": 33065, "stars": [], "topic_id": 34736, "date_created": 1305579470.012624, "message": "Financial Applications\n-----------------------------\n\n* Ability to analyze multiple overlapping time series\n* Different number of points in different time series - some might have a different update frequency, different time series might have different lengths\n* In the context of time series evolving means -> definition of 30 days forward changes every day. (different begin and end date every day). So 30 day window might mean May 1 - May 30 with a curve 0.1,0.1,0.2 .... next days curve will be May 2 - May 31 and values might be different too (0.3,0.2,0.1....). In essence even if the dates, the number of values in the curve, the dtypes, and the values in the curve vary they could be describing the same underlying entity.\n* Think of the curve as a vector rather than raster and write analysis code on top of this abstraction. This avoids code trying to fill gaps and align discrete data points - [Peter Wang]\n* Streaming data might have different sizes\n* Streaming data might have multiple ticks at the same time instant.\n* SQL/other querying on top of numpy would be nice.\n* indexing on top of numpy would be nice.\n* group by, aggregation operators on labels. \n* logical columns which are expressions on existing columns. (generator arrays)\n\nEric - Recurring issues\n-----------------------\n* need for labelled arrays\n* fit interpolation into arrays. so that when you ask for arr[2.3] it gives you a result \n* do arithmetic and functions on named columns. ", "group_id": 6727, "id": 1074790}, {"user_id": 33189, "stars": [], "topic_id": 34736, "date_created": 1305788573.798748, "message": "I just wrote up a short blog post about the summit, including a link to today's podcast: http://blog.fperez.org/2011/05/austin-trip-ipython-at-tacc-and.html.", "group_id": 6727, "id": 1109048}, {"user_id": 10411, "stars": [], "topic_id": 34736, "date_created": 1305789375.4686191, "message": "inSCIght episode about the summit is also up at http://inscight.org/2011/05/18/episode_13/", "group_id": 6727, "id": 1109100}, {"user_id": 33183, "stars": [], "topic_id": 34736, "date_created": 1306249910.663167, "message": "Nifty: http://www.haskell.org/haskellwiki/Numeric_Haskell:_A_Repa_Tutorial  Anyone seen or used this?", "group_id": 6727, "id": 1166879}, {"user_id": 33189, "stars": [], "topic_id": 34736, "date_created": 1306348039.230921, "message": "From my reading of it, though, the syntax for indexing is very verbose and cumbersome.  I can tell from practical experience that would be a pretty big issue for many users...  In real world data analysis, you write a *lot* of indexing code, and having to index things like", "group_id": 6727, "id": 1182052}, {"user_id": 33189, "stars": [], "topic_id": 34736, "date_created": 1306348066.5574369, "message": "x ! (Z :. 2) just to get x[2] looks pretty awful to me...", "group_id": 6727, "id": 1182059}, {"user_id": 33189, "stars": [], "topic_id": 34736, "date_created": 1306347809.4676261, "message": "Interesting, thanks for the link!  I should really learn some Haskell one of these days, I still struggle with the syntax...", "group_id": 6727, "id": 1182002}]