Alternatives To MS Excel?

[Cav]

I love the Microsoft software, but it’s very, very slow and cumbersome with big data and I’m at the point now where it’s becoming unworkable.

I can’t program databases and I really don’t have the time or inclination to learn.

Anyone know of any alternatives to MS Excel?

Is there a faster spreadsheet application out there that will handle hundreds of thousands of rows, is good with formulas and has pivot table or similar functionality. No bells and whistles required.

Thanks.

[thehorsesmouth]

Have you tried SPSS Statistics?

http://www-03.ibm.com/software/products … statistics

There’s a link to a free trial down that page and you should be able import your Excel file to see how it runs.

[wilsonl]

I know you’ve said you don’t want to use a DB Cav but something like Microsoft SQL server (which is free in express edition) wouldn’t be much more difficult to use for the basic requirements you’d have than a new Excel like application.

and it would be far more beneficial long term.

I’d be happy to help out with anything you needed.

Lee

[Cav]

Thanks for the replies.

Both look very complicated for a simpleton like me.

I just want to open my CSV’s, run formulas from them, then output and run lookup functions on related pivot tables.

Excel does this perfectly, but it’s a complete bugger with big files. Slow, cumbersome, unstable and destroys system memory.

I’ve got a good spec machine (16gb quad core i7) and I can’t believe there’s not a more efficient spreadsheet solution out there. But I can’t find one….

[Racing Daily]

There’s always Open Office which is freeware. The spreadsheet app is mostly the same as MS, but no doubt it will be less bloated than the Office alternative

[The Young Fella]

OpenOffice has declined badly in recent versions and seems more buggy and clunky than ever. When I bought a new laptop, I thought I’d be a cheapskate and get by with the OpenOffice suite instead of MS Office. A week of crashes and reduced functionality later and I’d scrapped that plan!

I used to use Lotus SmartSuite, but IBM support for that product ended in 2013 when they stopped marketing it. You can download it free as abandonware now, with third-party patches available to make it compatible with Windows 8. It’s a bit of a fuss to install on a new machine but always seemed lightweight.

[cormack15]

Cav – I’ve tried using Access to both store and analyse the data but the analysis part is much more cumbersome in Access and you can’t easily test and play with the data in the same way as Excel (at least I cant). I’m currently looking at storing the data in Access and using Excel to do the analytic by pulling in only the data I stipulate. ‘Looking at’ being the operative phrase.
The other alternative is to strip the stuff that may be adding less value from an analytic viewpoint and/or using different data sets for different analysis.

[Cav]

Thanks, I’ll take a look at Open Office, more in hope than expectation.

I find Access good for joining data files, but with my limited skills not a patch on Excel for analysis.

I’ve been Googling all day, couldn’t find anything, well not anything that can handle mass calculation without sending the system into meltdown.

Pretty surprising in this age of "Big Data".

Soldier on with Excel I guess.

[Prufrock]

Are you familiar with "R", Cav?

I dabbled with it briefly, didn’t really take to it, but people I know who also use Excel swear by it. It is undoubtedly MUCH quicker with big data.

[Marginal Value]

If you are determined to stick with a spreadsheet I can only echo Corm’s comments about making the spreadsheet smaller (in effect). By all means keep your Grand Spreadsheet with its hundreds of thousands of records. But it might be worth experimenting with creating several subset spreadsheets each with only one year’s data or just 20,000 records, running your calculations and analysis on each set, and then bringing the results together. You might find that Excel doing ten small jobs is much quicker than doing one job with ten times as much data. You do have the extra work of bringing the ten results together, but I guess that can be automated in a spreadsheet of only ten records to aggregate and analyse.

I used OpenOffice for quite a while, but a few years ago the developers, Oracle Corporation, decided to discontinue support for it. It has recently resurfaced as Apache OpenOffice but has had to play catch-up after a two or three year pause in development. I switched to LibreOffice which is run by a non-profit orgainization called The Document Fountation, whose mission is to produce open source free software adhering to Open Document Standards. LibreOffice and OpenOffice are very similar since they emerged from the same basic product, but I think that LibreOffice is better developed and better documented. There are far more free manuals, tutorials and YouTube videos for LibreOffice.

Although I use the spreadsheet (Calc) quite a lot, I also use the database (Base) for most of the data analysis. I use the HSQL database scheme out of the many options available. You can design and write Queries using three different methods that have different levels of assistance depending on whether you are a novice or an expert. You can design a Query with a high level of computer assistance and not have to write any SQL code. But once you get a bit of experience you can look behind the computer-assisted Query at the SQL code that is produced, which is not “all Greek to me” but surprisingly understandable using mostly normal English words. Many Queries can be created just by copying and pasting the SQL code and making little edits to change which data-fields you want to include, the criteria for selecting the records to use, the order in which to present the results, etc. You can, of course, use calculations in the Queries, such as turning beaten lengths into time difference according to the distance of the race, the going and the pace.

The biggest boon to using databases – faster search and query times – comes with splitting the data up into related data files or tables. Spreadsheets store data in two dimensions (rows and columns) and you can fake a bit of multidimensionalism by using look-up tables etc, but relational databases are the real deal for speed; in data entry, data retrieval and data analysis; and they minimise the amount of data that needs to be processed in each query. There are online tutorials, manuals, videos, or good old-fashioned, face-to-face college courses about relational databases, and, to top it all, helpful people like wilsonl on this forum. It might be worth the investment of time in the long run.

[quadrilla]

I’ve been using Qlikview for a few year’s now. You cannot get anything that’s faster – your full data is in memory. It’s FREE and really easy to use.

http://www.racing-tipster.com/wp-content/uploads/2015/01/SOUTHWELL-AW.pdf

Backing two runners is the relentless pursuit of value. Backing each way is a shortcut to the poor house. Only 7% make a long term profit.

[Cav]

Thanks again for the replies.

I tried OpenOffice. My largest spreadsheets take 25-30 seconds to open in Excel. Five minutes had passed and Calc was still crunching away trying to open my workbook, so that was hastily discarded.

I’ve tried R, Simon. Spent a good few hours calculating P-Values for a set of predictor variables I was using, along the lines of a tutorial written by Alun Owen that appeared in Smartsig some years ago. Can’t say the output told me anything about the data I didn’t know before the R analysis. The most statistically significant predictor variables were pretty much covered by the market. I didn’t find it the most user friendly of packages either.

Very good info in your post MV. I’ll also take a look at Qlikview when I get a chance.

[GeorgeJ]

Cavelino

Like you, I use Excel (in my case Excel 2007), to hold my race data and as the basis for my race analysis application.

How fast an Excel book processes depends on a number of factors – size, obviously, but also the computer on which it is being run and the coding of the book.

My main application is for analysing Flat handicaps and it includes three years full data and the data for the current year, so is obviously at its largest on 31 December, when it contains virtually four years data.

At 31 December 2014 the application had 206,233 rows each with 27 columns, which may well of course be much smaller than yours.

I run it on a PC with an i5 2320 processor, 8.0Gb of RAM (two x 4.0GB) with Windows 7, 64bit.

Opening the application takes about five seconds. If I then add data for a race to be analysed, and “start” the application, the time it takes to generate the analysis depends in major part on the number of runners, but to establish a benchmark as I have experimented with different PCs, versions of Windows and amounts of RAM I have used the race on 29 July 2014 won by Sennockian Star. It had 18 runners and I use that solely because it is typical of the sort of race I analyse.

Over the weekend I tested the application on three different machines, the time taken to produce the analysis once the book was open being as follows:

my main PC as specified above: 1 minute 56 secs.

on my back up PC, with a Pentium G2020, 4Gb of RAM (one chip), and Windows 7 32 bit: 25 minutes

on my son’s older PC, with a Pentium E5700, 4Gb of RAM (2 x 2Gb) and Windows 7 64 bit, 2 minutes 30 seconds.

I was very surprised by the second result, and although it is possible that having the 4 Gb of RAM in two chips rather than one was material I concluded that the more likely difference was having the 32 bit version of Windows on the machine.

Yesterday I upgraded it to the 64 bit version (which, with reloading MS Office and other bits and bobs took me about three hours). But worthwhile, as the time for processing the Sennockian Star race was down to 2 minutes 16 seconds.

The results now make sense, with the difference between my upgraded back up PC and my son’s trivial and probably reflecting the newer processor in the former, with the reduced time on my main machine reflecting the faster i5 processor and, possibly, more RAM. I suspect switching from the traditional hard drive to a solid state one would speed up significantly the processing time on any of the three, but I am not up to doing that myself and testing the effect of a solid state drive will have to await my next machine.

Allowing about five seconds for the application to open, and about two further minutes to analysis a current race or a past one (current ones take a few extra seconds because of the need to copy and paste the race data from the Post’s website), is for me very acceptable (whereas the 25 minutes the process took under Windows 32 was entirely unacceptable). Even with a faster processor – currently I would opt for an i7 – and a solid state drive, I’d be surprised (but very pleased) if the time was down more than about another 30 seconds.

You’ll hopefully be able to judge how my application and PC’s spec. compare to yours, but in the light of my experience if you are running on a Windows 7 32 bit machine I’d guess you would find things run much faster if you upgrade to the 64 bit version. I would add that I am wholly unclear why that is the case as such reading as I’ve done didn’t lead me to expect it, but empirically I am wholly convinced, given the initial comparison between my back up machine and my son’s with, other than in terms of the version of Windows 7, a slightly inferior spec.

[Author]

Posts