SQL scripted data correction

Using SQL 2012 I perform a manual process each month cleaning up erroneous data captured by the system. Yes it would be better to stop the data at the source but that's harder than it sounds.
I would like to script this entire process so I can run it either with a single call or even automate the running so that it happens before the SSRS report is generated. But I cant figure out how to put all the Hard pieces together with what I have now.

An example from todays work, I receive a trend for the month with unexplained spikes in the data. I note the date and run the following:

Declare @StDate DateTime
Declare @EndDate DateTime
Set @StDate = '6/06/2015 00:00:00.000' --  <<<<<<<<< SET BEGIN DATE
--Set @EndDate = DateAdd(month,1,@StDate) -- One Month
SET @EndDate =  DateAdd(day,1,@StDate) --One Day
Select dateadd(day,(datediff(day,0,TIMEandDATE)),0) as ReadingDate, TimeandDate,Value, id, tagid, _COMMENT
From MyTable
where 
TagID = 3004
and value > 0.06 -- <<<<< This varies depending on what I determine is a reasonable value based on before and after values
and TimeAndDate >= @StDate
and TimeAndDate < @EndDate

Select allOpen in new window

The results look like this:

2015-06-06 00:00:00.000	2015-06-06 20:00:00.953	0.0713222	8253107	3004
2015-06-06 00:00:00.000	2015-06-06 20:04:59.963	0.07127935	8253134	3004
2015-06-06 00:00:00.000	2015-06-06 20:09:59.947	0.07303441	8253161	3004
2015-06-06 00:00:00.000	2015-06-06 20:30:00.117	0.07020926	8253269	3004
2015-06-06 00:00:00.000	2015-06-06 20:35:00.107	0.1464455	8253296	3004 <<<<< Invalid Data Captured
2015-06-06 00:00:00.000	2015-06-06 20:40:00.193	0.07341963	8253323	3004
2015-06-06 00:00:00.000	2015-06-06 20:50:00.287	0.07256353	8253377	3004
2015-06-06 00:00:00.000	2015-06-06 21:50:00.563	0.07243514	8253701	3004
2015-06-06 00:00:00.000	2015-06-06 22:10:00.280	0.07252073	8253809	3004

Select allOpen in new window

So I find the ID number of the bad record and run this:

Update MyTable 
Set Value = 0.073 , -- <<<<<<<<<<<<<<<<<   SET CALCULATED (Reasonable)VALUE HERE
_COMMENT = LTRIM(ISNULL(_COMMENT,''+'Orig Value '+ CAST(Value AS VARCHAR(20)))) -- Places old value into comment field
where id = 8253296 ----  <<<<<<<<<<<<<<<<<<<<<<  ENTER ID OF INVALID DATA HERE
GO

Select allOpen in new window

Which then produces this:

2015-06-06 00:00:00.000	2015-06-06 20:00:00.953	0.0713222	8253107	3004	NULL
2015-06-06 00:00:00.000	2015-06-06 20:04:59.963	0.07127935	8253134	3004	NULL
2015-06-06 00:00:00.000	2015-06-06 20:09:59.947	0.07303441	8253161	3004	NULL
2015-06-06 00:00:00.000	2015-06-06 20:15:00.040	0.06931031	8253188	3004	NULL
2015-06-06 00:00:00.000	2015-06-06 20:20:00.037	0.06853986	8253215	3004	NULL
2015-06-06 00:00:00.000	2015-06-06 20:25:00.130	0.06926751	8253242	3004	NULL
2015-06-06 00:00:00.000	2015-06-06 20:30:00.117	0.07020926	8253269	3004	NULL
2015-06-06 00:00:00.000	2015-06-06 20:35:00.107	0.073		8253296	3004	Orig Value 0.146445 <<< CORRECTED VALUE
2015-06-06 00:00:00.000	2015-06-06 20:40:00.193	0.07341963	8253323	3004	NULL
2015-06-06 00:00:00.000	2015-06-06 20:45:00.183	0.06935316	8253350	3004	NULL
2015-06-06 00:00:00.000	2015-06-06 20:50:00.287	0.07256353	8253377	3004	NULL
2015-06-06 00:00:00.000	2015-06-06 20:55:00.267	0.06931031	8253404	3004	NULL
2015-06-06 00:00:00.000	2015-06-06 21:00:00.253	0.06776935	8253431	3004	NULL
2015-06-06 00:00:00.000	2015-06-06 21:05:00.350	0.06717008	8253458	3004	NULL

Select allOpen in new window

The issue I would like to solve is that the value I enter here is an approximation of the values before and after. Just had to be close, we are looking for a trend not specific numbers.
How would I automatically calculate a value to place in that single record and how would I do that for a months worth of data?
I would need to run through the months values looking for a number that is WAY higher than its predecessor and WAY higher than its successor. Then figure out what value to place there that is a reasonable mid value for the direction the data is trending.

Does this make sense? it does in my head .:)