Creating new array based on first element of each item in current array

Question

Please forgive me if my terminology isn't correct - I'm new to this!

I'm trying to isolate certain entries based on the value of the first element in each item in an array and perform some operations on certain columns. Here's an example of the type of data I'm working with:

[[1, 99, 400],
[1, 95, 200],
[2, 92, 100],
[1, 85, 500],
[2, 88, 300]]

I need to figure out the means of columns 2 and 3 using a for loop, if statement, and arithmetic, respectively, for each condition (reflected in column 1, where values are either 1 or 2).

I'm trying to split the above array into two separate arrays for each condition, and then taking the mean of those columns using numpy.mean. Here's what I want the lists to look like:

cond1 = [[1, 99, 400], [1, 95, 200], [1, 85, 500]]
cond2 = [[2, 92, 100], [2, 88, 300]]

I'm stuck on how to separate these conditions into two new arrays based on the first element. Here's the furthest I've gotten after googling about slicing.. but I'm stuck!

for x in stim:
    if stim_acc[0]==1.0:
        np.where(stim_acc[0] = 1.0)
        cond1 = [[s,a,rt] for s, a, rt in zip(stim, acc, mrt)]
        print(cond1)

regarding stim, acc, and mrt:

I have a several column-long set of data from which I isolated the stimulus (now stim), accuracy (now acc), and mean reaction time (now mrt) entries into a new list (the first in this post). I did this like so:

stim = data[:,1]
 acc = data[:,3]
 mrt = data[:,4]

 stim_acc = [[s, a, rt] for s, a, rt in zip(stim, acc, mrt)]
 print(stim_acc)

I've preemptively named the new list stim_acc because i foresaw this turning into a list calculating the accuracy for each condition over a loop.

Thank you, any help is greatly appreciated.

I'm not sure I understand what stim, acc, mrt are meant to be. Can you paste the full code snipet? — Neil
– Neil, Commented Oct 14, 2019 at 22:01
Just added an explanation for stim, acc, and mrt - let me know if you need more info! — microcastle
– microcastle, Commented Oct 14, 2019 at 22:10

Mert Köklü · Accepted Answer · 2019-10-14 22:01:16Z

1

As i understand, you want to split your list to two list by index[0]'s value. If it is 1 then add to cond1's list, else add to cond2's list. You can implement this using this code:

cond1=[]
cond2=[]
for item in listoflists:
    if item[0] == 1:
        cond1.append(item)
    else 
        cond2.append(item)

answered Oct 14, 2019 at 22:01

Mert Köklü

2,2512 gold badges19 silver badges21 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Mohsen_Fatemi · Accepted Answer · 2019-10-14 22:10:25Z

0

use pandas groupby :

>>> a = [[1, 99, 400],
 [1, 95, 200],
 [2, 92, 100],
 [1, 85, 500],
 [2, 88, 300]]
>>> df = pd.DataFrame(a)
>>> df
   0   1    2
0  1  99  400
1  1  95  200
2  2  92  100
3  1  85  500
4  2  88  300
>>> data = df.groupby([0])
>>> cond = data.groups
>>> df.loc[cond[1]]
   0   1    2
0  1  99  400
1  1  95  200
3  1  85  500
>>> df.loc[cond[2]]
   0   1    2
2  2  92  100
4  2  88  300

answered Oct 14, 2019 at 22:10

Mohsen_Fatemi

3,4212 gold badges21 silver badges32 bronze badges

Comments

Stuart · Accepted Answer · 2019-10-14 22:38:18Z

0

Using a simple loop, you can display the means like this:

data=[[1, 99, 400],
[1, 95, 200],
[2, 92, 100],
[1, 85, 500],
[2, 88, 300]]

groups = [], []
for row in data:
    groups[row[0]-1].append(row)

for group in groups:
    print(np.mean(group, axis=0))  # means of each column by group

But it is often better to use a pandas dataframe for this type of task

df = pd.DataFrame(data, columns=["stim", "acc", "mrt"])
for value in df.stim.unique():
    print(df[df.stim == value].mean())

or

for i, group in df.groupby("stim"):
    print(i, group.mean())

(I'm assuming you want the means of each column within each group.)

answered Oct 14, 2019 at 22:38

Stuart

9,8771 gold badge24 silver badges34 bronze badges

Collectives™ on Stack Overflow

Creating new array based on first element of each item in current array

3 Answers 3

Comments

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related