First, we’ll pull our normal libraries for working on projects like this.

import pandas as pd             # for the data.
import numpy as np              # for a NaN type
import matplotlib.pyplot as plt # For plotting, and some customization of plots.
import seaborn as sns           # For pretty plots.

# Fix the size of the graphs
sns.set(rc={"figure.figsize":(11, 8)})

Our data source is still going to be the Official Wiki and we’ll be pulling from the Loot Page - which has already started being updated for Season Two. From my previous post, the class zebra is still the best attribute to target to pull the data into a usable Data Frame.

# game taken down:
# url = "https://thecyclefrontier.wiki/wiki/Loot"
url = 'https://archive.ph/HZfyf'
site = pd.read_html(url, attrs={"class":"zebra"})[0]
site.head(15)

	Image	Name	Rarity	Weight	K-Marks	K-Marks / Weight	Personal Quarters	Campaigns	Jobs	Printing
0	NaN	Flawed Veltecite	Common	3.0	150 K-Marks 2 Reputation	50 K-Marks / Weight 0.67 Reputation / Weight	Yes x11	Yes x10	Yes	NaN
1	150.00	K-Marks	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
2	2.00	Reputation	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
3	50.00	K-Marks / Weight	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
4	0.67	Reputation / Weight	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
5	NaN	Cloudy Veltecite	Uncommon	3.0	570 K-Marks 6 Reputation	190 K-Marks / Weight 2 Reputation / Weight	Yes x8	Yes x14	Yes	Yes
6	570.00	K-Marks	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
7	6.00	Reputation	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
8	190.00	K-Marks / Weight	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
9	2.00	Reputation / Weight	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
10	NaN	Clear Veltecite	Rare	3.0	854 K-Marks 9 Reputation	285 K-Marks / Weight 3 Reputation / Weight	Yes x11	Yes x25	Yes	Yes
11	854.00	K-Marks	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
12	9.00	Reputation	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
13	285.00	K-Marks / Weight	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
14	3.00	Reputation / Weight	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN

Again, from previous experience, it is best to create a copy of columns that we actually need.

lootSubset = site[['Image', 'Name', 'Rarity', 'Personal Quarters', 'Campaigns', 'Jobs', 'Printing']].copy()
lootSubset.head()

	Image	Name	Rarity	Personal Quarters	Campaigns	Jobs	Printing
0	NaN	Flawed Veltecite	Common	Yes x11	Yes x10	Yes	NaN
1	150.00	K-Marks	NaN	NaN	NaN	NaN	NaN
2	2.00	Reputation	NaN	NaN	NaN	NaN	NaN
3	50.00	K-Marks / Weight	NaN	NaN	NaN	NaN	NaN
4	0.67	Reputation / Weight	NaN	NaN	NaN	NaN	NaN

Looking at our data, we already have a problem: the Printing column has columns NaN values insted of no. The blank values got converted to NaN when it was imported. While I’m sure there is a way to correct this on import, this isn’t how I’m going to do it. What I’m going to do is find the indexes which contain Yes and then update the other columns. You might be thinking - correctly - that we could simply find the indexes of the other values and then update those. Unfortunately, this is harder and I’ll show you what I mean.

If we check the values, we’ll see there is this nan value which doens’t look like the normal np.NaN value.

lootSubset.Printing.unique(), lootSubset.Printing[0]

(array([nan, 'Yes'], dtype=object), nan)

And, if we try to do a logical comparison with the literal value from the data it fails.

lootSubset.Printing.iloc[:5] == lootSubset.Printing[0]

0    False
1    False
2    False
3    False
4    False
Name: Printing, dtype: bool

Here we so the problem with objects and probably the pointers underneath. We could fix this but it’s much easier to reverse what we’re detecting so I’m doing that instead.

# That nan value is weird; will do the opposite
filterIndex = lootSubset.Printing == "Yes"
lootSubset[~filterIndex].head()

	Image	Name	Rarity	Personal Quarters	Campaigns	Jobs	Printing
0	NaN	Flawed Veltecite	Common	Yes x11	Yes x10	Yes	NaN
1	150.00	K-Marks	NaN	NaN	NaN	NaN	NaN
2	2.00	Reputation	NaN	NaN	NaN	NaN	NaN
3	50.00	K-Marks / Weight	NaN	NaN	NaN	NaN	NaN
4	0.67	Reputation / Weight	NaN	NaN	NaN	NaN	NaN

I’m not going to spend much time explaining this part; please see the previous posts about how this works.

# Correct it
lootSubset.loc[~filterIndex, "Printing"] = "No"
lootSubset.head(9)

	Image	Name	Rarity	Personal Quarters	Campaigns	Jobs	Printing
0	NaN	Flawed Veltecite	Common	Yes x11	Yes x10	Yes	No
1	150.00	K-Marks	NaN	NaN	NaN	NaN	No
2	2.00	Reputation	NaN	NaN	NaN	NaN	No
3	50.00	K-Marks / Weight	NaN	NaN	NaN	NaN	No
4	0.67	Reputation / Weight	NaN	NaN	NaN	NaN	No
5	NaN	Cloudy Veltecite	Uncommon	Yes x8	Yes x14	Yes	Yes
6	570.00	K-Marks	NaN	NaN	NaN	NaN	No
7	6.00	Reputation	NaN	NaN	NaN	NaN	No
8	190.00	K-Marks / Weight	NaN	NaN	NaN	NaN	No

Looking at the rows, we can see the data we’re after - the Name of the Loot - is in the Name column. Luckily, like the previous post about data extraction, these are in multiples of 5 insted of 4 like before. So, we can simply borrow the same code from before and update the range.

    # Change range to 5 instead of 4
    index = range( 0, len(lootSubset) - 4, 5)
    offset = np.array([1, 2, 3, 4])

tmp = index[0] + offset
lootSubset.iloc[tmp]

	Image	Name	Rarity	Personal Quarters	Campaigns	Jobs	Printing
1	150.00	K-Marks	NaN	NaN	NaN	NaN	No
2	2.00	Reputation	NaN	NaN	NaN	NaN	No
3	50.00	K-Marks / Weight	NaN	NaN	NaN	NaN	No
4	0.67	Reputation / Weight	NaN	NaN	NaN	NaN	No

lootSubset.iloc[index].tail()

	Image	Name	Rarity	Personal Quarters	Campaigns	Jobs	Printing
445	NaN	Hardened Metals	Common	Yes x17	Yes x2	Yes	Yes
450	NaN	Compound Sheets	Common	Yes x18	Yes x36	Yes	No
455	NaN	Print Resin	Uncommon	Yes x44	Yes x2	NaN	Yes
460	NaN	Salvaged Insulation	Common	Yes x38	Yes x30	NaN	No
465	NaN	Aluminum Scrap	Common	Yes x28	NaN	Yes	Yes

Now we’ll build the loop for the iteration like before.

# this is how we'll iterate; proof it works.
# refer the .head() call to see this is true.
for i in index[:3]:
    # we'll need to shift this to index 1 since 0 contains information we want.
    aLoot = lootSubset.iloc[i, 1]
    print(f'{aLoot} is at index {i}')

Flawed Veltecite is at index 0
Cloudy Veltecite is at index 5
Clear Veltecite is at index 10

Now we need a column to dump the correct values into which is going to be called Loot. We’ll fill these with np.NaN so we can repeat the same trick as the previous post.

lootSubset = lootSubset.assign(
    Loot = np.NaN
)
lootSubset.head()

	Image	Name	Rarity	Personal Quarters	Campaigns	Jobs	Printing	Loot
0	NaN	Flawed Veltecite	Common	Yes x11	Yes x10	Yes	No	NaN
1	150.00	K-Marks	NaN	NaN	NaN	NaN	No	NaN
2	2.00	Reputation	NaN	NaN	NaN	NaN	No	NaN
3	50.00	K-Marks / Weight	NaN	NaN	NaN	NaN	No	NaN
4	0.67	Reputation / Weight	NaN	NaN	NaN	NaN	No	NaN

So, pull the name of of the loot per subset and then update it like before.

for i in index:
    # Correct Loot column
    aLoot = lootSubset.iloc[i, 1]
    indexes = i + offset
    lootSubset.iloc[ indexes, 7 ] = aLoot

lootSubset.head(15)

	Image	Name	Rarity	Personal Quarters	Campaigns	Jobs	Printing	Loot
0	NaN	Flawed Veltecite	Common	Yes x11	Yes x10	Yes	No	NaN
1	150.00	K-Marks	NaN	NaN	NaN	NaN	No	Flawed Veltecite
2	2.00	Reputation	NaN	NaN	NaN	NaN	No	Flawed Veltecite
3	50.00	K-Marks / Weight	NaN	NaN	NaN	NaN	No	Flawed Veltecite
4	0.67	Reputation / Weight	NaN	NaN	NaN	NaN	No	Flawed Veltecite
5	NaN	Cloudy Veltecite	Uncommon	Yes x8	Yes x14	Yes	Yes	NaN
6	570.00	K-Marks	NaN	NaN	NaN	NaN	No	Cloudy Veltecite
7	6.00	Reputation	NaN	NaN	NaN	NaN	No	Cloudy Veltecite
8	190.00	K-Marks / Weight	NaN	NaN	NaN	NaN	No	Cloudy Veltecite
9	2.00	Reputation / Weight	NaN	NaN	NaN	NaN	No	Cloudy Veltecite
10	NaN	Clear Veltecite	Rare	Yes x11	Yes x25	Yes	Yes	NaN
11	854.00	K-Marks	NaN	NaN	NaN	NaN	No	Clear Veltecite
12	9.00	Reputation	NaN	NaN	NaN	NaN	No	Clear Veltecite
13	285.00	K-Marks / Weight	NaN	NaN	NaN	NaN	No	Clear Veltecite
14	3.00	Reputation / Weight	NaN	NaN	NaN	NaN	No	Clear Veltecite

If we play through the old trick of dropping the Loot columns now we’ll lose valuable data. So, we’ll need a new trick to keep that data. Thankfully, this is a problem I’d already solved at a previous time: we’ll simply fill the values. As a word of caution, filling values can be dangerous to an analysis so if you’re doing this then make sure it will not have a negative impact. In this instance, I’m duplicating the data so that it’s not lost at all. Even then, we’ll need to be cautious in the future about how the data is used.

There are a few techniques for filling in missing values and one such technique is called Fill Forward. What this does is take the values in a column, take that value and then simply inserts it down the rows where it finds NAs. This is exactly what we’re after - but only for those specific columns.

# fix the middle by filling down
tmp = lootSubset.iloc[:, 1:7]
tmp = tmp.fillna(method="ffill")
tmp.head(15)

	Name	Rarity	Personal Quarters	Campaigns	Jobs	Printing
0	Flawed Veltecite	Common	Yes x11	Yes x10	Yes	No
1	K-Marks	Common	Yes x11	Yes x10	Yes	No
2	Reputation	Common	Yes x11	Yes x10	Yes	No
3	K-Marks / Weight	Common	Yes x11	Yes x10	Yes	No
4	Reputation / Weight	Common	Yes x11	Yes x10	Yes	No
5	Cloudy Veltecite	Uncommon	Yes x8	Yes x14	Yes	Yes
6	K-Marks	Uncommon	Yes x8	Yes x14	Yes	No
7	Reputation	Uncommon	Yes x8	Yes x14	Yes	No
8	K-Marks / Weight	Uncommon	Yes x8	Yes x14	Yes	No
9	Reputation / Weight	Uncommon	Yes x8	Yes x14	Yes	No
10	Clear Veltecite	Rare	Yes x11	Yes x25	Yes	Yes
11	K-Marks	Rare	Yes x11	Yes x25	Yes	No
12	Reputation	Rare	Yes x11	Yes x25	Yes	No
13	K-Marks / Weight	Rare	Yes x11	Yes x25	Yes	No
14	Reputation / Weight	Rare	Yes x11	Yes x25	Yes	No

Then, we’ll take these values and simply insert them into the real data frame where we want them.

lootSubset.iloc[:, 1:7] = tmp
lootSubset.head(15)

	Image	Name	Rarity	Personal Quarters	Campaigns	Jobs	Printing	Loot
0	NaN	Flawed Veltecite	Common	Yes x11	Yes x10	Yes	No	NaN
1	150.00	K-Marks	Common	Yes x11	Yes x10	Yes	No	Flawed Veltecite
2	2.00	Reputation	Common	Yes x11	Yes x10	Yes	No	Flawed Veltecite
3	50.00	K-Marks / Weight	Common	Yes x11	Yes x10	Yes	No	Flawed Veltecite
4	0.67	Reputation / Weight	Common	Yes x11	Yes x10	Yes	No	Flawed Veltecite
5	NaN	Cloudy Veltecite	Uncommon	Yes x8	Yes x14	Yes	Yes	NaN
6	570.00	K-Marks	Uncommon	Yes x8	Yes x14	Yes	No	Cloudy Veltecite
7	6.00	Reputation	Uncommon	Yes x8	Yes x14	Yes	No	Cloudy Veltecite
8	190.00	K-Marks / Weight	Uncommon	Yes x8	Yes x14	Yes	No	Cloudy Veltecite
9	2.00	Reputation / Weight	Uncommon	Yes x8	Yes x14	Yes	No	Cloudy Veltecite
10	NaN	Clear Veltecite	Rare	Yes x11	Yes x25	Yes	Yes	NaN
11	854.00	K-Marks	Rare	Yes x11	Yes x25	Yes	No	Clear Veltecite
12	9.00	Reputation	Rare	Yes x11	Yes x25	Yes	No	Clear Veltecite
13	285.00	K-Marks / Weight	Rare	Yes x11	Yes x25	Yes	No	Clear Veltecite
14	3.00	Reputation / Weight	Rare	Yes x11	Yes x25	Yes	No	Clear Veltecite

And, now we’ll pull the same trick and delete the rows with np.NaN.

cutNA = lootSubset.Loot.isna()
lootData = lootSubset[ ~cutNA ]
lootData

	Image	Name	Rarity	Personal Quarters	Campaigns	Jobs	Printing	Loot
1	150.00	K-Marks	Common	Yes x11	Yes x10	Yes	No	Flawed Veltecite
2	2.00	Reputation	Common	Yes x11	Yes x10	Yes	No	Flawed Veltecite
3	50.00	K-Marks / Weight	Common	Yes x11	Yes x10	Yes	No	Flawed Veltecite
4	0.67	Reputation / Weight	Common	Yes x11	Yes x10	Yes	No	Flawed Veltecite
6	570.00	K-Marks	Uncommon	Yes x8	Yes x14	Yes	No	Cloudy Veltecite
...	...	...	...	...	...	...	...	...
464	0.60	Reputation / Weight	Common	Yes x38	Yes x30	Yes	No	Salvaged Insulation
466	506.00	K-Marks	Common	Yes x28	Yes x30	Yes	No	Aluminum Scrap
467	5.00	Reputation	Common	Yes x28	Yes x30	Yes	No	Aluminum Scrap
468	51.00	K-Marks / Weight	Common	Yes x28	Yes x30	Yes	No	Aluminum Scrap
469	0.50	Reputation / Weight	Common	Yes x28	Yes x30	Yes	No	Aluminum Scrap

376 rows × 8 columns

Almost there! The column name Image is not what that should be called so we’ll update that to Unit like we had in the prevous post.

lootData = lootData.rename(columns={'Image':'Unit'})
lootData

	Unit	Name	Rarity	Personal Quarters	Campaigns	Jobs	Printing	Loot
1	150.00	K-Marks	Common	Yes x11	Yes x10	Yes	No	Flawed Veltecite
2	2.00	Reputation	Common	Yes x11	Yes x10	Yes	No	Flawed Veltecite
3	50.00	K-Marks / Weight	Common	Yes x11	Yes x10	Yes	No	Flawed Veltecite
4	0.67	Reputation / Weight	Common	Yes x11	Yes x10	Yes	No	Flawed Veltecite
6	570.00	K-Marks	Uncommon	Yes x8	Yes x14	Yes	No	Cloudy Veltecite
...	...	...	...	...	...	...	...	...
464	0.60	Reputation / Weight	Common	Yes x38	Yes x30	Yes	No	Salvaged Insulation
466	506.00	K-Marks	Common	Yes x28	Yes x30	Yes	No	Aluminum Scrap
467	5.00	Reputation	Common	Yes x28	Yes x30	Yes	No	Aluminum Scrap
468	51.00	K-Marks / Weight	Common	Yes x28	Yes x30	Yes	No	Aluminum Scrap
469	0.50	Reputation / Weight	Common	Yes x28	Yes x30	Yes	No	Aluminum Scrap

376 rows × 8 columns

Just one more problem I’d like to correct before we do some fun questions at the end. Pandas has also copied the idea of Categoreies from R. And, it allows us to set the order of the values as well. We’re going to use this in a minute but it’s also a nice to have.

Changing the type is just as easy as you’d think:

lootData.Rarity.astype('category')

1        Common
2        Common
3        Common
4        Common
6      Uncommon
         ...   
464      Common
466      Common
467      Common
468      Common
469      Common
Name: Rarity, Length: 376, dtype: category
Categories (6, object): ['Common', 'Epic', 'Exotic', 'Legendary', 'Rare', 'Uncommon']

You can see that now there is a list of the possible Categories below the values: Categories (6, object): ['Common', 'Epic', 'Exotic', 'Legendary', 'Rare', 'Uncommon']. However, the order matters and so we cannot leave this how it is; Uncommon is clealy not larger in type then Epic. The proper way to do this would be to declare the order - which is what is next. We’ll use pd.Categorical() to convert them and set the order of the Categories.

lootData['Rarity'] = pd.Categorical(lootData.Rarity, categories = ['Common', 'Uncommon', 'Rare', 'Epic', 'Exotic', 'Legendary'])
pd.Categorical(lootData.Rarity, categories = ['Common', 'Uncommon', 'Rare', 'Epic', 'Exotic', 'Legendary'])

['Common', 'Common', 'Common', 'Common', 'Uncommon', ..., 'Common', 'Common', 'Common', 'Common', 'Common']
Length: 376
Categories (6, object): ['Common', 'Uncommon', 'Rare', 'Epic', 'Exotic', 'Legendary']

There we go! See that the order has now been fixed: Categories (6, object): ['Common', 'Uncommon', 'Rare', 'Epic', 'Exotic', 'Legendary']. Here I would like to note that I did take liberty to set Legendary above Exotic since the wiki as them mixed and the actual games sorting puts Legendary above Exotic and so I did the same.

Conclusion and Some Questions

And, that’s how you pull data from a website! Now for some interesting questions. Looking at the data we have here, there are two questions that popped out at me: 1. If you have to pick up only two items, what would they be per Rarity? 2. Given the sell values per category, is the relationship linear as the the category rises?

What Are The Best Two Items To Sell, Per Rarity?

This is a pretty simple ask. We just have to remember that we’re working with Tidy Data and we’ll want to filter the Name column for K-Marks before doing anything else. Next, we’ll do a groupby() for the Rarity and then pull out the Units column since that has the values we’re after. Luckily, pandas already has a function to get the largest values .largest() which also works as an aggregation function for groupby().

tmp = lootData.query("Name == 'K-Marks'").groupby("Rarity")['Unit'].nlargest(2)
tmp

Rarity        
Common     336       900.0
           111       760.0
Uncommon   356      3417.0
           181      1709.0
Rare       371     11533.0
           366      5126.0
Epic       121     20183.0
           396     17300.0
Exotic     401    129746.0
           106     77848.0
Legendary  441    518985.0
           431    116772.0
Name: Unit, dtype: float64

So, we have our values but we don’t have any way to identify them since we cannot include the Name column since then it becomes a Data Frame and .nlargest() doesn’t work. If we look closely, the old indexes have been carried over and are included in our results so we’ll need to give ourselves access to them. We’re going to do this with .reset_index() since that will push those indexes into the Data Frame.

tmp.reset_index()

	Rarity	level_1	Unit
0	Common	336	900.0
1	Common	111	760.0
2	Uncommon	356	3417.0
3	Uncommon	181	1709.0
4	Rare	371	11533.0
5	Rare	366	5126.0
6	Epic	121	20183.0
7	Epic	396	17300.0
8	Exotic	401	129746.0
9	Exotic	106	77848.0
10	Legendary	441	518985.0
11	Legendary	431	116772.0

# Keep those index numbers
tmp = tmp.reset_index()

The index column is called level_1 which we’re going to simply insert into the original data.

# Best items to sell per rarity:
lootData.loc[ tmp['level_1'].tolist() ].sort_values("Rarity", ascending=False)[['Unit', 'Name', 'Rarity', 'Loot']]

	Unit	Name	Rarity	Loot
441	518985.0	K-Marks	Legendary	Alpha Crusher Head
431	116772.0	K-Marks	Legendary	Savage Marauder Head
401	129746.0	K-Marks	Exotic	Alpha Crusher Heart
106	77848.0	K-Marks	Exotic	Progenitor Slag
121	20183.0	K-Marks	Epic	NiC Oil Cannister
396	17300.0	K-Marks	Epic	Crusher Flesh
371	11533.0	K-Marks	Rare	Crusher Hide
366	5126.0	K-Marks	Rare	Mature Rattler Eyes
356	3417.0	K-Marks	Uncommon	Hardened Bone Plates
181	1709.0	K-Marks	Uncommon	Derelict Explosives
336	900.0	K-Marks	Common	Toxic Glands
111	760.0	K-Marks	Common	Progenitor Composite

There we go! Well, except that this doesn’t account for weight and what we really want is to carry the most value per weight since we’re limited in the game by the backpack size. Let’s do the same but for the per weight value instead.

tmp = lootData.query("Name == 'K-Marks / Weight'").groupby("Rarity")['Unit'].nlargest(2).reset_index()
lootData.loc[ tmp['level_1'].tolist() ].sort_values("Rarity", ascending=False)[['Unit', 'Name', 'Rarity', 'Loot']]

	Unit	Name	Rarity	Loot
443	17300.0	K-Marks / Weight	Legendary	Alpha Crusher Head
433	5839.0	K-Marks / Weight	Legendary	Savage Marauder Head
108	6487.0	K-Marks / Weight	Exotic	Progenitor Slag
403	6487.0	K-Marks / Weight	Exotic	Alpha Crusher Heart
123	4037.0	K-Marks / Weight	Epic	NiC Oil Cannister
118	3844.0	K-Marks / Weight	Epic	Letium Clot
373	1153.0	K-Marks / Weight	Rare	Crusher Hide
368	1025.0	K-Marks / Weight	Rare	Mature Rattler Eyes
458	570.0	K-Marks / Weight	Uncommon	Print Resin
358	380.0	K-Marks / Weight	Uncommon	Hardened Bone Plates
148	1000.0	K-Marks / Weight	Common	Old Currency
283	506.0	K-Marks / Weight	Common	Nutritional Bar

And, now we see there are some important changes. The values in the Common Rarity are totally different!

Are the Rarity Values Linear?

We’ll aggregate over all the values in a category and then simply plot them.

lootData.query("Name == 'K-Marks'").groupby("Rarity")['Unit'].mean().plot()
plt.title("Mean K-Marks Per Rarity", size=15, fontweight='bold')

Text(0.5, 1.0, 'Mean K-Marks Per Rarity')

Wait, what? This looks terrible! Everything from Common to Epic is worth nothing compared to Exotic and Legendary. Maybe we’re dealing with some outlier problems? Let’s check the median just in case.

lootData.query("Name == 'K-Marks'").groupby("Rarity")['Unit'].median().plot()
plt.title("Median K-Marks Per Rarity", size=15, fontweight='bold')

Text(0.5, 1.0, 'Median K-Marks Per Rarity')

This is even worse! Legendary loot is definitely being affected by outliers but Exotic is so far and above better than everything else. Maybe we’re looking at the wrong values? Maybe we’re making the same mistake and we need to account for the per weight?

lootData.query("Name == 'K-Marks / Weight'").groupby("Rarity")['Unit'].median().plot()
plt.title("Median K-Marks/Weight Per Rarity", size=15, fontweight='bold')

Text(0.5, 1.0, 'Median K-Marks/Weight Per Rarity')

I mean it’s better but not really. What is in this category?

lootData.query("Rarity == 'Exotic'").query("Name == 'K-Marks'")

	Unit	Name	Rarity	Personal Quarters	Campaigns	Jobs	Printing	Loot
91	8650.0	K-Marks	Exotic	Yes x4	Yes x9	Yes	No	Charged Tharis Iron Ingot
106	77848.0	K-Marks	Exotic	Yes x6	Yes x9	Yes	No	Progenitor Slag
401	129746.0	K-Marks	Exotic	Yes x6	Yes x1	Yes	No	Alpha Crusher Heart

This category makes no sense and I don’t know why it exists. I’m not sure they understand what this category is for either looking at what is in here. But, that will be it for now.