How to find first local maximum for every group?

Question

Suppose I have a Pandas data frame as follows:

Test Parameter Value

X1     0        0.033285423511615113
X1     1        0.78790279861666179
X1     2        0.79136989638378297
X1     3        0.80063190842016707
X1     4        0.7884653622402551
X1     5        0.78561849214309198...
...
X1     22       22: 0.82241991278171311...
...
X2 ...

I'd like to get the row with Parameter value 3. That is the row with the last increasing value before the first drop. Notice that later on we might have higher values (eg row 22). Essentially, I'm trying to get the "last" number before the "first" decrease value.

Also note that there are multiple Tests, so I probably need to do something like:

myDF.groupby("Test").Something

Answer 1

Coldspeed nearly has it, to get only the first group you can use cumprod, or similar e.g.

In [11]: df[((df.Value.diff().fillna(1) > 0).cumprod()) == 1].tail(1)
Out[11]:
  Test  Parameter     Value
3   X1          3  0.800632

The trick being:

In [12]: (df.Value.diff().fillna(1) > 0)
Out[12]:
0     True
1     True
2     True
3     True
4    False
5    False
6     True
Name: Value, dtype: bool

In [13]: (df.Value.diff().fillna(1) > 0).cumprod()
Out[13]:
0    1
1    1
2    1
3    1
4    0
5    0
6    0
Name: Value, dtype: int64

---

Note: My df is this:

In [21]: df
Out[21]:
  Test  Parameter     Value
0   X1          0  0.033285
1   X1          1  0.787903
2   X1          2  0.791370
3   X1          3  0.800632
4   X1          4  0.788465
5   X1          5  0.785618
6   X1         22  0.822420