Chapter 16 — interactive

The Digital Twin

The case of Mia: compute in your browser how her twin depends on who she is talking to.

A digital twin is never the picture of a person alone — it is the picture of a relationship. Here you see it in black and white: the same Mia shows a different mix of the five symbiont roles with each person. The Python code from the book runs straight in your browser; you don't need to install anything.

Bee — the creative Ant — the doer Butterfly — the shaper Capybara — the harmoniser Leech — the taker

Pyodide is loading… (10–20 seconds the first time)

The workshop

Four cells, one shared memory. What you compute above is still available below — so run them from top to bottom.

Step 1Mia's messages

Five relationships, a few synthetic messages — only Mia's side counts for her twin.

mias_chats = {
    "Sister": [
        "Omg I've got the most awesome idea for Dad's birthday 🎉",
        "Let's build him a scavenger hunt through the whole flat lol",
        "or better: a home-made video as a surprise!",
        "come on, be creative with me 😄 it'll turn out beautiful",
        "you handle the decorations, I'll take on the tasks around it",
    ],
    "Teacher": [
        "Dear Mr Berger, please find attached the outline of my presentation.",
        "The sources are numbered, the deadline is scheduled for Friday as agreed.",
        "Does this schedule suit you? I'll follow your guidelines.",
        "Thanks very much, that's very kind of you. happy to.",
    ],
    "Friend": [
        "hey, you were so quiet today 🫂",
        "want to talk? no stress, i'm here",
        "we'll get through this together, nice and calm",
        "reach out whenever you like ❤️",
        "and then we'll do something beautiful together, let's get out 😄",
    ],
    "Mother": [
        "Mum will you drive me to the station at 8?",
        "and will you lend me another 20 francs? I'll pay it back promise",
        "can you bring me that please, still need something for school",
        "thanks you're the best 😘 🙈 I'll clean up later promise",
        "oh and I've got a sweet idea for the weekend together 😄",
    ],
    "Employer": [
        "Good morning, I'll take the early shift as agreed.",
        "Delivery is stocked, till reconciled.",
        "should I come in earlier tomorrow? I'm flexible.",
        "thanks, happy to, see you later!",
    ],
}
print("Messages loaded for", len(mias_chats), "relationships.")

Expected: “Messages loaded for 5 relationships.”

Step 2The word lists

For each role a small list of telltale words. (Roles, not labels for people — see Chapter 6.)

signalwords = {
    "Bee":       ["idea", "creative", "let's", "build", "or better",
                  "what if", "awesome", "lol", "😄", "🎉", "surprise"],
    "Ant":       ["outline", "plan", "sources", "deadline", "schedule",
                  "as agreed", "shift", "reconciled",
                  "stocked", "guidelines", "flexible", "tasks", "tidy"],
    "Butterfly": ["beautiful", "design", "layout", "colour", "picture", "shape",
                  "style", "video", "great", "pretty"],
    "Capybara":  ["no stress", "i'm here", "together", "calm", "talk",
                  "🫂", "❤️", "all good", "thanks", "best", "happy to"],
    "Leech":     ["can you", "lend me", "drive me", "give me", "for me",
                  "still need", "you do", "😘", "🙈", "promise"],
}
print("Word lists defined for", len(signalwords), "roles.")

Defines signalwords.

Step 3Compute the shares

For each relationship: count the hits per role and convert to per cent. The aggregate is the mean over all relationships.

roles = list(signalwords)

def count_hits(text):
    text = text.lower()
    t = {r: 0 for r in roles}
    for r, words in signalwords.items():
        for w in words:
            t[r] += text.count(w)
    return t

def shares(messages):
    total = {r: 0 for r in roles}
    for m in messages:
        for r, n in count_hits(m).items():
            total[r] += n
    overall = sum(total.values())
    return {r: round(100 * total[r] / overall) if overall else 0 for r in roles}

per_relationship = {who: shares(chat) for who, chat in mias_chats.items()}
aggregate = {r: round(sum(p[r] for p in per_relationship.values()) / len(per_relationship))
             for r in roles}

header = "Relationship".ljust(13) + "".join(r[:4].rjust(7) for r in roles)
print(header); print("-" * len(header))
for who, p in per_relationship.items():
    print(who.ljust(13) + "".join(f"{p[r]:6d}%" for r in roles))
print("-" * len(header))
print("AGGREGATE".ljust(13) + "".join(f"{aggregate[r]:6d}%" for r in roles) + "   <- Mia's twin")

Needs Steps 1 and 2.

Step 4Mia's five faces — the chart

One bar per relationship, the aggregate below. The first time, matplotlib loads briefly.

import matplotlib
matplotlib.use("agg")
import matplotlib.pyplot as plt, io, base64

colour = {"Bee":"#E0A82E","Ant":"#6B7A8F","Butterfly":"#C77DA0",
          "Capybara":"#6F9E78","Leech":"#B5524C"}
rows = list(per_relationship.items()) + [("Mia's twin (aggregate)", aggregate)]

fig, ax = plt.subplots(figsize=(8.6, 4.6))
ys = [7, 6, 5, 4, 3, 1.3][:len(rows)]
for (label, p), y in zip(rows, ys):
    left = 0
    bold = label.startswith("Mia's twin")
    for r in roles:
        ax.barh(y, p[r], left=left, height=0.72, color=colour[r],
                edgecolor="white", linewidth=1.1)
        if p[r] >= 9:
            ax.text(left + p[r]/2, y, str(p[r]), ha="center", va="center",
                    color="white", fontsize=8.4, fontweight="bold")
        left += p[r]
    ax.text(-1.5, y, label, ha="right", va="center", fontsize=10.4,
            fontweight="bold" if bold else "normal")

ax.plot([-0.2,100.2],[2.15,2.15], color="#CDCDC7", lw=1.1, ls=(0,(4,3)))
ax.set_xlim(0,100); ax.set_ylim(0.6, 7.9); ax.set_yticks([])
for s in ["top","right","left"]: ax.spines[s].set_visible(False)
ax.set_xticks(range(0,101,25)); ax.set_xticklabels(["0","25","50","75","100%"], fontsize=9)
ax.tick_params(axis="x", length=0)
ax.set_title("Mia's five faces - and her twin", fontsize=13, fontweight="bold")

buf = io.BytesIO()
plt.savefig(buf, format="png", dpi=150, bbox_inches="tight", facecolor="white")
plt.close(fig)
_chart_b64 = base64.b64encode(buf.getvalue()).decode()
print("Chart created.")

Needs Step 3.

Try it out

Change one of Mia's messages in Step 1 — to her mother, say — so that she gives more than she takes. Run Steps 3 and 4 again: how far does the red Leech share drop, and how does the aggregate shift?

Or enter your own (voluntarily shared, anonymised) chats in Step 1. You test two lessons from the book for yourself: the twin depends on the other person, and a good twin is an aggregate — but even the aggregate hides how different a person is from one relationship to the next.