Lets Build a Tank Parameter/Chemistry Database to Better Understand What Works

[JohnnyTabasco]

What about trace elements?
Maybe an implementation of ICP tests as well.

 

[aarbutina]

@aarbutina presented this idea of course, but I was thinking if someone like R2R got involved, this could be a massive database unlike any other. Simple web based entry, where every user name gets a form that can be updated and maintained with data and media. Imagine a dynamic, detailed, living build thread for thousands of aquariums. It could be searchable by cross referencing dozens of parameters such as equipment, livestock, parameters, etc.

I think I just drooled a little.

I was also thinking that this endeavor could greatly benefit from some official support from the R2R staff, if only to drive participation (maybe a link on the front page or some kind of “call it arms”).

My thinking would be that a web based form while pretty would require a bit more upkeep than a simple spread sheet. if a new parameter is added to be tracked I could see it taking a bit of work getting that implemented as opposed to just adding a new column to a data table. But on the other hand it would be very slick looking and might drive additional participation.

One other thing to keep in mind is that it may not be the best idea for people to be constantly up dating with their latest parameters. I think a snap shot of the tank at a particular moment in time would lend it self more to a better picture of what works and what doesn’t. That being said it could be possible to build a dynamic data base that can be queried and analyzed (as a snap shot). It would be a great way to trade trend in how people are keeping their tanks as a community.

I like you’re thinking @NY_Caveman keep it up.

 

[aarbutina]

@Catfish2 and @JohnnyTabasco i would support the addition of trace Elements but the only problem I could foresee if the number of people who actually test for these. The integration of IPC test results could certainly help but while many people already do this testing I am not so sure people who don’t do this testing would be willing to participate if they had to spend money to do so. Can any one say sponsorship?

 

[Gareth elliott]

You could use google docs, have anyone interested enter their data into an online spreadsheet after permissions have been granted. Then revoke after they have entered them, if what you want is collection of tank snap shots at 1 moment in time.

As far as tank chemistry, with everyone using different hobby grade tests. How do you plan to remove any results that are test or human error?

 

[Shameless_Dood]

One of the things I do as a hobby / gonna be a second career is software development. It'd be a fun little exercise for me to create the db and a simple web page where people can enter the info.

I'm just curious how often you want to collect data, how long. Etc etc.

 

[aarbutina]

You could use google docs, have anyone interested enter their data into an online spreadsheet after permissions have been granted. Then revoke after they have entered them, if what you want is collection of tank snap shots at 1 moment in time.

As far as tank chemistry, with everyone using different hobby grade tests. How do you plan to remove any results that are test or human error?

Good suggestions.

The hope would be this a large enough sample size that the variation in the data due to human error or testing differences would be minimize. Unfortunately this kind of error is just going to exist. Also with a large enough sample size gross errors should be kicked out as outliers.

 

[aarbutina]

One of the things I do as a hobby / gonna be a second career is software development. It'd be a fun little exercise for me to create the db and a simple web page where people can enter the info.

I'm just curious how often you want to collect data, how long. Etc etc.

Now that is what I like to hear!

As for the intervals I think it will depend if we can actually get this thing off the ground and get people involved. I think the first milestone would be just populating the table once, just one big data dump. If we find that participation is high and interest continues then we could think about how do we turn this into a dynamic system that is constantly being updated.

 

[Gareth elliott]

Before you start collecting data. Might also help to get those who want to be involved to state tank age and what they test for. I own most of the tests you list but i rarely test more than alkalinity, tbh. How many test their tanks for all of these parameters on older systems? I dont know the answer to that lol. But might want a controlled sample size in respect to tank age. Otherwise might bias the results to a specific range that might not be as useful as intended.
Example 20 members give their data but 17 have had their systems running 6-9 months, 1 at 24m 2 at 36m.

Perhaps when starting open spots based on age 10 for x months, Another 10 for y months, etc.

Then when have enough data open it for more data points.

Also how do you control for incomplete data? I personally have no desire to buy a par meter. I know a ball park for my lights from others and lux conversions but not the actual numbers on my system. My par numbers would not be as accurate as someone with a par meter. Do i leave blank? Not included in sample? Enter my less accurate par numbers?

I applaud your idea, so dont take any of these questions as me trying to dissuade you. Just wondering on the mechanics.

 

[aarbutina]

@Gareth elliott these are all good questions. And I totally hear you that it is likely that folks with well established tanks may have dropped of with a large number of the tests that they may have been performing when the tank was first started off. It may be a situation where those folks may want to do a one time test so they can include their numbers (but if we could figure out how to get a large amount of participation in IPC testing that could help with that).

As of other incomplete data like in your situation PAR numbers, it would be possible to perform data analysis in multiple ways. For instance first look at all the data with PAR values excluded. Then you can perform the analysis again but this time filter out the values that don't have PAR values associate with it. Depending on what the analysis shows we can figure out how to fill in the missing data points if they are proving to be critical.