Publication-quality figures — LaTeX labels, colormaps, DPI

5.4.19 · Coding › Scientific Computing (Python)

1. DPI — "resolution" ka matlab asal mein kya hai

YEH KYUN MATTER KARTA HAI? Journals ek column width inches mein specify karte hain (jaise 3.5 in) AUR ek minimum DPI (aksar line art ke liye 300, fine detail ke liye 600). Agar aap sirf pixel size set karo toh aapka koi control nahi hoga ki page par text kitna bada lagta hai.

Worked example Single column ke liye sizing

Journal chahta hai 3.5 in column, minimum 300 dpi.

fig, ax = plt.subplots(figsize=(3.5, 2.6))   # physical size
fig.savefig("fig.png", dpi=300)              # 1050 x 780 px

figsize pehle kyun? Kyunki font sizes points mein hote hain (1 pt = 1/72 in). figsize set karne se inch canvas fix ho jaata hai, toh ek 9-pt label paper par apna real size rakhta hai chahe DPI kuch bhi ho.

2. LaTeX / math labels

Definition True LaTeX enable karna

plt.rcParams.update({
    "text.usetex": True,
    "font.family": "serif",
    "font.serif": ["Computer Modern Roman"],
})
ax.set_xlabel(r"$\lambda$ (nm)")

Aage ka r ek raw string banata hai taaki Python backslash \lambda ko nahi khaata.

RAW STRINGS KYUN? Ek normal string mein "\lambda" → \l ek jaani-pehchaani escape nahi hai lekin "\n" hai ek newline. r"..." use karne se guarantee hai ki backslashes LaTeX tak intact pahunchte hain.

Worked example Units ke saath labeled axis

ax.set_ylabel(r"Intensity $I/I_0$")
ax.set_title(r"$E = mc^2$")

$...$ kyun? Dollar signs math mode mein switch karte hain bilkul LaTeX ki tarah; unke bahar text upright/roman hota hai.

3. Colormaps — perceptual uniformity

Worked example Heatmap + colorbar sahi se karna

im = ax.imshow(Z, cmap="viridis", origin="lower",
               extent=[x0, x1, y0, y1], aspect="auto")
cbar = fig.colorbar(im, ax=ax)
cbar.set_label(r"$|\psi|^2$")

colorbar kyun? Bina labeled scale ke ek color map unreadable hoti hai — colorbar hi color ka y-axis hai.

Figure — Publication-quality figures — LaTeX labels, colormaps, DPI

4. Sab kuch mila kar (puri recipe)

Formula Reproducible figure block

plt.rcParams.update({
    "font.size": 9,            # journal body ≈ 8–10 pt
    "axes.labelsize": 9,
    "savefig.dpi": 300,
    "savefig.bbox": "tight",   # trims white margins
    "text.usetex": True,
})
fig, ax = plt.subplots(figsize=(3.5, 2.6))
...
fig.savefig("fig.pdf")         # vector → infinite DPI for line art

Line plots ke liye PDF kyun? Vector formats shapes store karte hain, pixels nahi, isliye DPI irrelevant hai — text aur lines kisi bhi zoom par razor sharp hote hain. PNG/TIFF high DPI par sirf raster images (photos, dense heatmaps) ke liye use karo.

Recall Feynman: 12-saal ke bachhe ko samjhao

Socho ek drawing kar rahe ho kisi real book mein daalne ke liye. DPI yeh hai ki tum ek inch mein kitne tiny dots thunste ho — bahut kam ho toh picture Minecraft jaisi blocky lagti hai. LaTeX labels tumhare picture ke words aur math ko book ke fancy fonts se match karate hain taaki woh out of place na lage. Colormaps woh rule hai jo numbers ko colors mein badalta hai — ek achha rule bade numbers ko step by step brighter paint karta hai, taaki tumhari aankh picture ko honestly padhe instead of ek flashy rainbow se cheat ho.

Common mistakes

Flashcards

3.5 in × 2.6 in figure jo dpi=300 par save ki gayi ho, uske kitne pixels hote hain?

1050 \times 780

px (har inch dimension ko 300 se multiply karo).

viridis ko jet se prefer kyun karein?

viridis perceptually uniform hai (monotonic brightness), grayscale-safe aur colorblind-friendly hai; jet false edges create karta hai aur grayscale mein fail ho jaata hai.

r"$\lambda$" mein aage ka r kya karta hai?

Ek raw string banata hai taaki Python backslash escapes interpret na kare; \lambda LaTeX tak intact pahunchta hai.

Sequential aur diverging colormaps mein kya fark hai?

Sequential low→high data ke liye (viridis); diverging aisi data ke liye jiska meaningful center/zero ho, center ko neutral color par map karta hai (coolwarm).

Line plots PDF mein kyun save karein PNG ki jagah?

PDF vector hai — pixels nahi shapes store karta hai, isliye kisi bhi zoom par sharp hota hai aur DPI irrelevant hai; PNG raster hai aur DPI par depend karta hai.

Figure ki physical printed size kya set karta hai?

figsize (inches mein), DPI nahi. DPI sirf uss size par sharpness set karta hai.

Font sizes se pehle figsize kab set karna chahiye?

Hamesha — fonts points mein hote hain (1/72 in), isliye inch canvas determine karta hai ki page par text kitna bada dikhega.

text.usetex=True ka kya kaam hai default mathtext se alag?

usetex rendering ek real LaTeX install ko deta hai (exact paper fonts); mathtext matplotlib ka built-in TeX-jaisa engine hai, koi external LaTeX nahi chahiye.

Connections

Matplotlib basics — figure and axes objects
Colormaps and color theory in visualization
LaTeX typesetting
Raster vs vector graphics
Dimensional analysis
Reproducible research and rcParams